Skip to main content

Full text of "A diagrammatic approach to categorification of quantum groups III"

See other formats

A diagrammatic approach to categorification of 

56 ; quantum groups III 


^ ■ Mikhail Khovanov and Aaron D. Lauda 


July 22, 2008 


Q^, We categorify the idempotented form of quantum sl(n) 

^ : Contents 

1 Introduction 


Q ■ 2 Graphical Interpretation of the bihnear form 

'^ ■ 2.1 Quantum groups 

-.^ . 2.2 Geometric interpretation of the bilinear form . 

(^ ■ 3 Graphical calculus for U categorification 

OO , 3.1 The 2-category W 

3.2 Spanning sets of HOMs in W 

3.3 Properties and symmetries of 2-category U 
/\f ' 3.4 Karoubi envelope, U, and 2-representations 

c^ • 3.5 Direct sum decompositions 

3.6 Kq{U) and homomorphism 7 

3.7 Idempotented rings 

3.8 Surjectivity of 7 

3.9 Injectivity of 7 in the nondegenerate case 

Categorification of U for s[„ 

4.1 Forms of quantum stn 

4.2 The 2-category W_^(s[„,) 

5 Iterated fiag varieties 

5.1 Cohomology of n-step varieties 

5.2 Graphical calculus for iterated flag varieties 

5.3 The 2-category Flag^ 

6 Representing W^ on the flag 2-category 

6.1 Defining the 2-functor F^v 

6.2 Checking the relations of W_ 

6.3 Equivariant representation 

6.4 Nondegeneracy of W_^(s[n) 



References 86 

1 Introduction 

In this paper we categorify the Beilinson-Lusztig-Macpherson idempotented modification 
\J{sln) of \Jq{sln) for any n, generalizing [211 [22] where such categorification was described 
for n = 2 and using constructions and resuhs of [Ml [IT] which contain a categorification 
of U~ for any Cartan datum. 

In [21] Lusztig associates a quantum group U to any root datum; the latter consists 
of a perfect pairing (, ) between two free abelian groups X and Y, embeddings of the set 
/ of simple roots into X,Y, and a bilinear form on Z[/] subject to certain compatibility 
and integrality conditions. Lusztig's definition is slightly different from the original ones 
due to Drinfeld [9] and Jimbo ^3\. Lusztig then modifies U to the nonunital ring U which 
contains a system of idempotents {1a} over all weights A G X as a substitute for the unit 

u = i.uu. 


In the sin case U was originally defined by Beilinson, Lusztig, and Macpherson [2j and 
then appeared in [23l [H] in greater generality. It is clear from Lusztig's work that the 
Q(g)-algebra U is natural for at least the following reasons: 

1. A U- module is the same as a U- module which has an integral weight decomposition. 
These modules are of prime importance in the representation theory of U. 

2. U has analogues of the comultiplication, the antipode, and other standard symmetries 
of U. 

3. U is a U-bimodule. 

4. The Peter- Weyl theorem and the theory of cells can be intrinsically stated in terms 
of the algebra U. 

5. U has an integral form _4U, a Z[g, g~^]-lattice closed under multiplication and co- 
multiplication. The integral form comes with a canonical basis B. Conjecturally, 
multiplication and comultiplication in this basis have coefficients in N[g, g~^] when 
the Cartan datum is symmetric. 

6. The braid group associated to the Cartan datum acts on U. 

Moreover, U appears throughout the categorification program for quantum groups. 
Representations of quantum groups that are known to have categorifications all have in- 
tegral weight decompositions, and thus automatically extend to representations of U. In 
most or all of these examples, see [31 U\ [101 EH [33], the weight decomposition of repre- 
sentations lifts to a direct sum decomposition of categories, so one obtains a categorifi- 
cation of the idempotent 1^ as the functor of projection onto the corresponding direct 
summand (the only possible exception is the categorification of tensor products via the 
affine Grassmannian O [5]). In the categorification of tensor powers of the fundamental 
Uq(sl2)-representation [3], [lO], each canonical basis element of U acts as an indecompos- 
able projective functor or as the zero functor. The idea that U rather than U should be 
categorified goes back to Crane and Frenkel [8]. 

U is generated by elements -EjIa, Filx, and 1a, where A G X is an element of the weight 
lattice and i is a simple root. We will often write E^i instead of Ei and E_i instead of Fj. 
We have 

E±ilx = lfj_E±ilx, 

where, in our notations, explained in Section [2?T| n = \±ix, and ix is the element of X 
associated to the simple root i. Algebra U is spanned by products 

Eilx := E±i^E±i^ . . . E±i^lx = l^E±i-^E±i^ . . . E±i^lx, 

where i = (±ii, . . . , ±im) is a signed sequence of simple roots, and // = A + ix- 
The integral form ^U C U is the Z[g, g~^]-algebra generated by divided powers 

Ei(a)i. = -p-pr-Ef Ia- 

Note that U can, alternatively, be viewed as a pre-additive category with objects A G X 
and morphisms from A to /i being I^UIa. Of course, any ring with a collection of mutually- 
orthogonal idempotents as a substitute for the unit element can be viewed as a pre-additive 
category and vice versa. From this perspective, though, we can expect the categorification 
of U to be a 2-category. 

In Section 13.11 we associate a 2-category W to a root datum. The objects of this 2- 
category are integral weights A G X, the morphisms from A to /i are finite formal sums 
of symbols Silx{t}, where i = {±ii, . . . , ±im) is a signed sequence of simple roots such 
that the left weight of the symbol is /i {Eilx = 1^-EjIa), and t G Z is a grading shift. 
When i consists of a single term, we get 1-morphisms Sj^ilx and S-ilx, which should be 
thought of as categorifying elements Eilx and FjIa of U, respectively. Grading shift {t} 
categorifies multiplication by g*. The one-morphism Silx '■ A — >fi should be thought of as 
a categorification of the element Eilx- When the sequence i is empty, we get the identity 
morphism 1a : A — >X, a categorification of the element 1a. 

Two-morphisms between Silx{t} and Sjlx{t'} are given by linear combinations of 
degree t — t' diagrams drawn on the strip IR x [0, 1] of the plane. The diagrams consist 

of immersed oriented one- manifolds, with every component labelled by a simple root, and 
dots placed on the components. Labels and orientations at the lower and upper endpoints 
of the one-manifold must match the sequences i and j , respectively. Integral weights label 
regions of the plane cut out by the one-manifold, with the rightmost region labelled A. Each 
diagram has an integer degree assigned to it. We work over a ground field k, and define 
a two-morphism between Silx{t} and Sjlx{t'} as a linear combination of such diagrams 
of degree t — t', with coefficients in k, modulo isotopies and a collection of very carefully 
chosen local relations. The set of 2-morphisms U{Silx{t} , Sjlx{t'}) is a k-vector space. 
We also form graded vector space 

iiOMu{Silx,Sjlx) := ^U{Silx{t},Sjlx). (1.1) 

Vertical composition of 2-morphisms is given by concatenation of diagrams, horizontal 
composition consists of placing diagrams next to each other. 

In each graded k-vector space B.OMu{£ilx, Sjlx) we construct a homogeneous spanning 
set Bijx which depends on extra choices. The Laurent power series in q, with the coefficient 
at q"^ equal to the number of spanning set elements of degree r, is proportional to suitably 
normalized inner product {Eilx, Ejlx), where the semilinear form (, ) is a mild modification 
of the Lusztig bilinear form on U. The proportionality coefficient vr depends only on the 
root datum. 

We say that our graphical calculus is nondegenerate for a given root datum and field k 
if for each i,j and A the homogeneous spanning set Bijx is a basis of the k-vector space 
}iOMu{Silx,Sjlx)- Nondegeneracy will be crucial for our categorification constructions. 

The 2-category U is k-additive, and we form its Karoubian envelope U, the smallest 
2-category which contains U and has splitting idempotents. Namely, for each A, /i G X, the 
category U{X, yu) of morphisms A — >fi is defined as the Karoubian envelope of the additive 
k-linear category W(A, /i). The split Grothendieck category Kq(U) is a pre-additive category 
with objects A, and the abelian group of morphisms from A to yU is the split Grothendieck 
group Ko(U{X, fij) of the additive category W(A, /i)). The grading shift functor on W(A, /i) 
turns Ko(U{X, fi)) into a Z;[g, g"^]-module. This module is free with the basis given by 
isomorphism classes of indecomposable objects of ZY(A,/i), up to grading shifts. The split 
Grothendieck category Kq(U) can also be viewed as a nonunital Z[g, g~^]-algebra with a 
collection of idempotents [1a] as a substitute for the unit element. 

In Section [3l6l we set up a Z[g, g^^]-algebra homomorphism 

which takes 1a to [1a] and E^lx to [^jIa], for any "divided power" signed sequence i. 
The main results of this paper are the following theorems. 

Theorem 1.1. The map 7 is surjective for any root datum and field k. 

Theorem 1.2. The map 7 is injective if the graphical calculus for the root datum and 
field k is nondegenerate. 

Theorem 1.3. The graphical calculus is nondegenerate for the root datum oi sin and any 
field k. 

The three theorems together immediately imply 

Proposition 1.4. The map 7 is an isomorphism for the root datum of sin and any field 


The last result establishes a canonical isomorphism 


and allows us to view W(s[„) as a categorification of U(s[„). 

Theorem 11.11 proved in Section I3.8[ follows from the results of J16l [T71 |2T] and basic 
properties of Grothendieck groups and idempotents. Theorem ll.2[ proved in Section 13.91 
follows from the nondegeneracy of the semilinear form on U and its pictorial interpretation 
explained in Section 12.21 To prove theorem 11.31 we construct a family of 2-representations 
of U and check that the elements of each spanning set Bijx act linearly independently on 
vector spaces in these 2-representations, implying nondegeneracy of the graphical calculus. 
Sections |1H6] are devoted to these constructions. 

Indecomposable one-morphisms, up to isomorphism and grading shifts, constitute a 
basis of Ko(U{sln)) = A^i^ln), which might potentially depend on the ground field k. The 
multiplication in this basis has coefficients in N[g, q~^]. It is an open problem whether this 
basis coincides with the Lusztig canonical basis of ^U(s[„). The answer is positive when 
n = 2, see [21]. 

Another major problem is to determine for which root data the graphical calculus is 
nondegenerate. Nondegeneracy immediately implies, via theorems 11.11 and II. 2[ that U 
categorifies U for a given root datum. 

We believe that U will prove ubiquitous in representation theory. This 2-category or 
its mild modifications is expected to act on parabolic-singular blocks of highest weight 
categories for sl^ in the context of categorification of sin representations [TJ [TOl [31], on 
derived categories of coherent sheaves on Kronheimer-Nakajima [20] and Nakajima [26] 
quiver varieties and on their Fukaya-Floer counterparts, on categories of modules over 
cyclotomic Hecke and degenerate Hecke algebras [Il[19], on categories of perverse sheaves in 
Zheng's categorifications of tensor products [33], on categories of modules over cyclotomic 
quotients of rings R{i^) in [13 [T7], on categories of matrix factorizations that appear in [18], 
Section 11], et cetera. A possible approach to proving that the calculus is nondegenerate 
for other root systems is to show that U acts on a sufficiently large 2-category and verify 
that the spanning set elements act linearly independently. It would also be interesting to 
relate our constructions with those of Rouquier [2Sl 130] • 

Categories of projective modules over rings R{i'), defined in [161 [H]) categorify U~ 
weight spaces. A subset of our defining local relations on two-morphisms gives the relations 

for rings R{v). This subset consists exactly of the relations whose diagrams have no critical 
points (U-turns) on strands and have all strand orientations going in the same direction. 
In other words, the relations on braid-like diagrams allow us to categorify U~, while the 
relations without these restrictions lead to a categorification of the entire U, at least in 
the sin case. Informally, the passage from a categorification of U^ to a categorification of 
U is analogous to generalizing from braids to tangles. 

Acknowledgments. M.K. was partially supported by the NSF grant DMS-0706924 
and, during the early stages of this work, by the Institute for Advanced Study. 

2 Graphical Interpretation of the bilinear form 

2.1 Quantum groups 

2.1.1 Algebras f and U 

We recall several definitions, following [21]. A Cartan datum (/, ■) consists of a finite set / 
and a symmetric Z- valued bilinear form on Z[/], subject to conditions 

• i -i e {2,4,6,. . .} ioT i e I, 

• dij := -2|f G {0, 1, 2, ... } for any t^jin I. 

Let qi = q'^, [a]i = q^'^ + gf"^ + ■ ■ ■ + q]''', [a]i\ = [a]i[a - l]i . . . [1]^. Denote by 'f the free 
associative algebra over (Q(g) with generators 6i, i & I, and introduce q-divided powers 
^i = ^?/Hi'- The algebra 'f is N[/]-graded, with 9i in degree i. The tensor square 'i0'i 
is an associative algebra with twisted multiplication 

(xi X2){x[ (g) X2) = g"'^^'''^i'xiXi 0:2X2 

for homogeneous Xi, 0:2, x[, x^- The assignment r(di) = di®\W®di extends to a unique 
algebra homomorphism r : 'f — >'f ® 'f . 

Algebra 'f carries a (Q(g)-bilinear form determined by the condition^j 

• (1,1) = 1, 

• (0„0,) = (5,,,(l-g2)-ifor^,je/, 

• (a;, yy') = {r{x),y (g) y') for x, y, y' G 'f , 

• {xx' , y) = {x ® x' , r{y)) for x, x', y G 'f . 

The bilinear form (, ) is symmetric. Its radical U is a two-sided ideal of 'f . The form (, ) 
descends to a nondegenerate form on the associative (Q(g)-algebra f = 'f/J. 

^Our bilinear form (,) corresponds to Lusztig's bilinear form {,}, see Lusztig [24, 1.2.10]. 

Quantum groups 

Theorem 2.1. The ideal 3 is generated by the elements 


over all i,j & I, i j^ j. 

It seems that the only known proof of this theorem, for a general Cartan datum, requires 
Lusztig's geometric realization of f via perverse sheaves. This proof is given in his book ^M 
Theorem 33.1.3]. Less sophisticated proofs exist when the Cartan datum is finite. 

We see that f is the quotient of 'f by the quantum Serre relations 

J2 (-l)"^f^^,# = 0. (2.1) 


Let ^f be the Z[g, g^^]-subalgebra of f generated by the divided powers Of" , over all 
i e I and a G N. 

A root datum of type (J, ■) consists of 

• free finitely generated abelian groups X,Y and a perfect pairing (, ) : F x X ^ Z; 

• inclusions I <Z X {i ^^ ix) and I dY {i y^ i) such that 

• {i,ix) ='^T-i= -d-ij for all i,j E I. 

This implies {i, ix) = 2 for all i. We write ix-, rather than Lusztig's i' , to denote the image 
of i in X. 

The quantum group U associated to a root datum as above is the unital associative 
Q(g)-algebra given by generators -Ej, Fj, K^ for i G / and ji eY , subject to the relations: 

i) Kq = 1, Kij,K^> = Kf,+^> for all /i, fi' G Y, 

ii) Kf,Ei = q'^^'''^^EiK^ for alH G /, /i G Y, 

iii) K^Fi = q-^^'''^^FiK^ for alH G /, /i G Y, 

iv) EiFj - FjEi = 6ij ^"ZH-'i ^ where K±i = K^a.imi, 

v) For all i ^ j 

E i-irEt^E,Ei'^ = and ^ (-1)^^^/"^^^.^^^ = 0- 

a+6=-(ijx>+l a,+b=-(i,jx)+'i- 

li f{6i) G 3 for a polynomial / in noncommutative variables 6i, i E I, then f{Ei) = 
and f{Fi) = in U. This gives a pair of injective algebra homomorphisms f ^ U. 

Quantum groups 8 

2.1.2 Some automorphisms of U 

Let " be the Q-hnear involution of Q(g) which maps q to q^^. U has the following standard 
algebra (anti) automorphisms. 

• The Q(g)-antilinear algebra involution ip: \J ^U given by 

^(E,) = E,, tlj{F,) = F„ ^P{K^) = K_^, ^(/x) = Mx) for / G (Q(g) and xeV. 

• The Q(g)-linear algebra involution uj: U ^U given by 

uj{Ei) = F,, uj{Fi) = E,, uj{K^) = K^^. 

• The Q(g)-linear algebra antiinvolution a:\J^ U°p given by 

a{Ei) = Ei, a{Fi) = F,, a{K^) = K^. 

• The Q(g)-linear algebra antiinvolution p: U ^ U°p given by 

p{Ei) = q.KiFi, p{Fi) = q^K.^Ei, p{K^) = K^. 

We denote by p the Q(g)-linear antiinvolution ippip: U -^ U°p. 

• The (Q(g)-antilinear antiautomorphism r : U ^ U°p defined as the composite r = ipp: 

T{Ei) = q^'K^,F„ t{F,) = q~'K,E„ r{K^) = K_^. 

2.1.3 U and the bilinear form 

The (Q(g)-algebra U is obtained from U by adjoining a collection of orthogonal idempotents 
1a for each A G X, 

UIa' = 5a,a'1a', (2.2) 

such that 

ir^lA = lA^M = 9^'''^^lA, F.Ia = 1a+^;,^., FaA = lA-.x^.- (2.3) 

The algebra U decomposes as direct sum of weight spaces 

U= Ia'UIa. 


We say that A, respectively A', is the right, respectively left, weight of a; G Ia'UIa. The 
algebra ^U is the Z[g, g~^]-subalgebra of U generated by products of divided powers E^ Ix 
and Fj 1a, and has a similar weight decomposition 

^U= 1a'UU)1a. 



Quantum groups 


The following identities hold in U and ^U: 




The (anti) automorphisms ip, u, a, p and r all naturally extend to U and ^U if we set 

^Ij{1x) = 1a, c^(1a) = 1-A, ^(1a) = 1-A, p(1a) = U, ^(Ia) = Ia- 

Taking direct sums of the induced maps on each summand Ia'UIa allows these maps to 
be extended to U. For example, the antiautomorphism r induces for each A and A' in X 
an isomorphism Ia'UIa -^ IaUIa'. Restricting to the Z[g, g~^]-subalgebra ^U and taking 
direct sums, we obtain an algebra antiautomorphism r: ^^U -^ ^U such that t{1x) = Ix, 

771(1)1 1 rpia) p(a)-| 

- 1 p('^) 

- ^X-aix-'^i 5 



a — b + {i, X) 


min(a,6) ^ 
^— n 

—a + b — {i, A) 

j^{a-t) p{b-t) ^ 
- i 


IxFilx+ix^ and r(lAFjlA+ 




Ix+ixEdx for all XeX. 

T{lx+ixEilx) = q'i 

The following result is taken from Lusztig [231 26.1.1], but our bilinear form is normal- 
ized slightly differently from his. 

Proposition 2.2. There exists a unique pairing (, ) : U x U ^ Q(q') with the properties: 
(i) (,) is bilinear, i.e., ifx,y) = f{x,y), (xjy) = f{x,y), for / G Q(g) and x,y e\J, 
(ii) (1aiX1a2 7 Ia'^I/Ia^) = for all x,y &\J unless Ai = A'^ and A2 = Ag, 
(iii) {ux,y) = {x,p{u)y) for -u G U and x, y G U, 

(iv) (xIa, x'lx) = {x, x') for all x, x' G f = U+ and all A (here (x, x') is as in Section [2.1.ip . 
(v) (x, y) = {y, x) for all x, y G U. 
Definition 2.3. Define a semilinear form (, ) : U x U ^ Q(q') by 

(x, y) := {ip{x), y) for all x, y G U. 
Proposition 2.4. The map (, ) : U x U ^ Q(q') has the following properties 
(i) (,) is semilinear, i.e., {fx,y) = f{x,y), {xJy) = f{x,y), for / G Q(g) and x,?/ G U, 
(ii) (IaiXIaj, 1a'Z/1a') = for all x, y G U unless Ai = A'^ and A2 = A' 
(iii) {ux, y) = (x, T{u)y) for m G U and x, y G U, 

*2 5 


Quantum groups 10 

(iv) {xlx,x'lx) = {'ip{x),x') for x, x' G f and A G X, 

(v) {x,y) = {ilj{y),^{x)) for all x,y etj. 

Proof. Immediate. D 

Proposition 2.5. The bilinear form (, ) and the semilinear form (, ) are both nondegenerate 
on U. 

Proof. Nondegeneracy of (, ) is implicit throughout [211 Chapter 26] and follows for instance 
from Theorem 26.3.1 of [24J. D 

The bilinear and semilinear forms restrict to pairings 

(, ) : ^U X ^U ^ Z[q, q~'], (, ) : ^U x ^U ^ Z[g, q-']. (2.8) 

2.1.4 Signed sequences 

Throughout the paper we write E^i for Ei and E^i for Ei and need notation for products of 
these elements. By a signed sequence i = (eiii, £2^2, • • • , ^mim), where ei, . . . ,em G {+, — } 
and ii, . . . ,im G /, we mean a finite sequence of elements of / with signs. We may also 
write i as eiii, £2^2, • • • , ^mim or even £:iZi£:2^2 • • • £mim- Let \i\ = Yl^=i ^k^k, viewed as an 
element of Z[/]. The length of i, denoted || i ||, is the number of elements in the sequence 
i [m in the above notations). Define SSeq to be the set of signed sequences. Let 

Ei := E^^i^E^^i2 . . . -E£,„j„ G U, (2.9) 

EiU ■■= Ee,,,E,,,, . . .^„„.„1a e U. (2.10) 

Recall that ix & X denotes the image of i under the embedding I ^ X. Let 

ix ■■= ieiii)x H h iemim)x G X. (2.11) 

Then Eil\ = Ix+i^Eilx, so that A, respectively A + ix, is the right, respectively left, 
weight oi Eilx- 

A sequence i is positive if all signs ei, . . . ,6^ = +■ We may write a positive sequence 
as (ii, . . . , im) or even ii . . . im- Denote by SSeq^ the set of all positive sequences. As in 
[16], for z/ G N[J], z/ = ^ z/j ■ i, denote by Seq(z/) the set of all sequences i = ii- ■ -im such 
that i appears Vi times in the sequence. There is an obvious bijection 

SSeq+ = W Seq(z/). (2.12) 

A sequence is negative if all signs Si, . . . .Sm = —■ For a positive sequence i denote by 
— i the corresponding negative sequence. For any signed sequence i denote by — i the 
sequence given by reversing all signs in i. We write concatenation of sequences i and j as 
ij, attaching ±2 to the left of i as ±i«, etc. 


Quantum groups 


By a divided powers signed sequence (dpss for short) we mean a sequence 

i = {sA'''\e2i^'\...,e^zt-^) (2.13) 

where e's and i's are as before and ai,...,am G {1,2, .. . }. Let \i\ = J2T=i^k<^kik G ^[-^] 
and define || i ||, the length of i, as X^fcLi ^k ^ N. For a dpss i let i be a signed sequence 
given by expanding i 

[Eiti, . . . , Eiti, £2^2, • • • , ^2^2, • • • , ^m^r 

I Ern.f'r 

with term ei^i repeating ai times, term £2^2 repeating 02 times, etc. 






the quantum divided powers of generators E'ej. Then 

-ei(«) -LA = -Egj 1a G U 


and has left weight A + eaix ■ More generally, for a dpss « 

E- ■= _£;("!)_£; ("2) 

ei«i £2*2 

&m I'm 





with left weight A + YlT=i ^rO'r{ir)x- 

Let SSeqd be the set of all dpss. Elements -EjIa, over all i G SSeqd and A G X, span 
^U as a Z[g, g~^]-module. For z/ G N[/], Seqd(i^) was defined in |i6] as the set of all 
expressions if . . . im such that J2T=i ^rir = ^- Each element of Seqd(z/) gives a positive 
divided power sequence, and 

SSeqd+ = W Seqd(z/). (2.18) 

The length || i || of a dpss is defined as the sum ai + 02 + ■ ■ ■ + a^ in the above notation. 
For the reader's convenience our notations for sequences are collected below: 

positive divided (_ 
powers sequences 
of weight V 

positive sequences "- 
of weight V 


-)■ positive divided power "- 

signed sequences 

^ Seq+ e. 

positive sequences 

-¥ divided powers 
signed sequences 

signed sequences 


Geometric interpretation of the bilinear form 


2.2 Geometric interpretation of the bilinear form 

Mark m points 1 x {0}, 2 x {0}, . . . , -m x {0} on the lower boundary R x {0} of the strip 
R X [0,1] and k points 1 x {1},2 x {1},...,A; x {1} on the upper boundary R x {1}. 
Assuming m + k is even, choose an immersion of ^^^^ strands into R x [0, 1] with these 
m + k points as the endpoints. Orient each strand and label it by an element of /. Then 
endpoints inherit orientations and labels from the strands: 

+ - - + 
— • • • •— k = A 



Orientation and labels at lower and upper endpoints define signed sequences i, j G SSeq. In 
the above example, i = {+i, —j, +i, —i) and j = {+i, — j, —i, +i). We consider immersions 
modulo boundary-preserving homotopies and call them pairings between sequences i and 
j, or simply (i, j)-pairings. 

Clearly, (i, j)-pairings are in a bijection with complete matchings of m + k points such 
that the two points in each matching pair share the same label and their orientations are 
compatible. Denote by p'{i,j) the set of all (i, j)-pairings. 

A minimal diagram D oi a (i, j)-pairing is a generic immersion that realizes the pairing 
such that strands have no self-intersections and any two strands intersect at most once. 
We consider minimal diagrams up to boundary-preserving isotopies. A (i,jf) -pairing has 
at least one minimal diagram, in the example below it has two: 

Here is an immersion which is not a minimal diagram 

Given A G X, color the regions of a minimal diagram D by elements of X such that the 
rightmost region is colored A, and the two regions between the two sides on an i-labelled 
strand differ by ix 

\ + ix 



Geometric interpretation of the bilinear form 


Isotope D so that the strands at each crossing are either both oriented up or down, and de- 
fine the degree deg{D, A) of D relative to A as the integer which is the sum of contributions 
from all crossings, local maxima and minima of D: 















Notice that 


,....=, i±(.A), 


In the simply-laced case Qi = q and c±i^x = 1 ± (?, A). 

The degrees of the other crossings are determined from the rules above 








^\\ ^ 

V J 



Since both of these crossings have degree zero, we will refer to them as balanced crossings. 

Proposition 2.6. deg(-D, A) depends only on A and on the (i, j)-pairing realized by D. 
Thus, deg{D, A) is independent of the choice of a minimal diagram for a pairing. 

Proof. Invariance under cancellation of U-turns can be shown as follows: 


/A + ix 



\ + ix 

^(1 + (z. A)) + '-^{l-{^,X + r^)) = = de^ 


Invariance of deg(D, A) under other isotopies of D is straightforward and is left to the 
reader. Any two minimal diagrams of the same (i, j)-pairing are related by a sequence of 
isotopies and "triple-crossing" moves for various orientations, see examples below. 

A ^ 

The invariance under these moves is manifestly obvious. 




Geometric interpretation of the bilinear form 


From now on, for each (i, j)-pairing we choose one minimal diagram D representing 
this pairing and denote by p{i,j) the set of these diagrams. Recall the pairings (, ) and 
(, ) on U from Section I2.1[ 

Theorem 2.7. For any i,j G SSeq and X, ft E X 

Dep(i,.i) i£i 



where Kj is the number of i-colored strands in D. 
For example, 


so that 


(i5;(_,,+,„+,,_.)lA,i?(-,,_.)lA) = (g^'---.^+--+..^-^- + g-+.^^--x) _^ _j^^ (2.23) 



(1 - q^? (1 - g.) 

The product term in (12.221) is independent of D since 2K,i is the number of times label 
i appears in the sequence ij. 

Proof. The first equality in (I2.22p is obvious since -EjIa is invariant under the involution 
tl). For i, J, A, /i as in Theorem 12.71 define 

{E,i.,E,i,)'=6,, y: ^'"^"''^ n ttz^ 


During the proof we view -EjIa and Ejl^ in {Eilx,Ejl^)' as formal symbols rather than 
elements of U, since we have not yet proved that (, )' descends to U. We want to show 

{Eilx,Ejl^) = {Eilx, Ejlf^i) 

for all i, j, A and /i. 

The nontrivial case is A = /x. We can also interpret the formula {Eilx,Ejlfj)' = for 
fj' y^ X via the sum of diagrams, since then the rightmost region must be colored by both 
A and fi, which is not allowed. 


Geometric interpretation of the bilinear form 


Notice that in the absence of (i,j) -pairings the right hand side of fl2.22p is zero, since 
there are no diagrams to sum over. If the left weights of -EjIa and Ejlx are different the 
inner product (-EjIa, Ej1\) = and (-EjIa, Ejlx)' = as well since the set of (i,j) -pairings 
is empty in this case. 

Assume that both i and j are positive. Then any minimal (i, j)-diagram can be 
isotoped to be braid-like, with the strands going upward: 

Moreover, in this case k = m, otherwise p{i,j) is empty. The sum on the right hand side 
of (12.221) describes the canonical bilinear form on U+ = f evaluated on Ei = Ei^ ■ ■ ■ Ei^ 
and Ej = Ej^ . . . E^-^ , see [23 EZl [HI [IH • 

In view of property (iv) of Lusztig's bilinear form, we obtain the following: 

Lemma 2.8. U i, j are positive, 

{Edx, EjUy = (Edx, EjU) (2.25) 

for any A G X. 

Thus, for positive i, j , the geometrically defined bilinear form (, )' coincides with (, ). 

Lemma 2.9. Equation (12.221) holds for the pair (±u, j) if and only if it holds for the pair 


Proof. Attaching a [/-turn gives a bijection between p{±ii,j) and p{i, =FU)- 




We have 

t ■ I 

I ■ I 

deg((j). A) = deg(D, A) + — (1 ± (z, /i T ix)) = deg(D, A) - — (1 ^ {^, /i)), 
where the additional term matches the power of q in the formula p{l^E±i) = q^"^ E^il 



Geometric interpretation of the bilinear form 


The previous lemma implies that it is enough to check the equality 

{Eax,Ejlxy=iEilx,EjU) (2.26) 

when j is the empty sequence. The next three lemmas show that 

(EilxAxY = (EdxAx) (2.27) 

for any i and A. 

Lemma 2.10. Equation (12.221) holds for all pairs ((— j)i,0) with i, j positive. 

Proof. Bending up — j by adding U-turns transforms the pair ((— j)i, 0) to {i,j) for which 
fl2:22|) holds (see Lemma El]). We have 

iE^-j)ilx, 1a) = q'^iEdx, Ejlx), (^(-,-)aA, U)' = q'^iEdx, E^U)', (2.28) 

where a is the power of q in the formula p{E_jlx+ix) = '^i.E^jlx+ix) = I'^^x+ixEj- □ 
Lemma 2.11. For i,jGl,i^ j 

{Ei'±i^ji"lx, 1a)' = {Ei'^i^ji"lx, 1a) (2.29) 

if and only if 

{Ei'^j±ii"lx, 1a)' = {Ei'^j^ii"lx, 1a). (2.30) 

Proof. Attach a crossing at the ±i =F j location to a diagram D in p{i' ±2=1= ji", 0). 

Tj ±i 

The resulting diagram D' is minimal if the ±i and =Fj strands of D do not intersect. 
Otherwise, it is not minimal, but the homotopy 

will make it minimal (the orientations in the picture are for the +i — j case) 
We get a bijection 




Geometric interpretation of the bilinear form 17 

„.. „s .e or . .a.a.„, .. a.. . ..e .. (X ^ ^ .. a.. .. 

{Ei'±i,pji"l\, 1a)' = (i?i'=pj±ij"lA, 1a)'- (2.32) 

Since Ei'^i^ji"lx = Ei'^jj^n'ilx in U, we see that 

{Ei'±i^ji"l\-, 1a) = {Ei'^j±ii"l\, 1a) (2.33) 

and the lemma follows. D 

Lemma 2.12. Assume {Eiii"l\, 1\)' = {Eiii"l\, 1\). Then 

(E,'+,_,,"1a, 1a)' = (^i'+.^.i"lA, 1a) (2.34) 

if and only if 

(-Ej'_j+ii"lA, 1a)' = (-Ei'_j+ij"lA, 1a)- (2.35) 

Proof. Decompose the diagrams in p{i' + i — ii", 0) into three classes 

(1) +i and —i strands do not intersect V J 

+i — i 

(2) +i and —i strands intersect 

+i —i 

(3) a strand connects +i and —i -/'"^ 

+i —i 
Likewise, decompose the diagrams in p{i' — i + ii", 0) into three classes: 

'1) —i and +i strands do not intersect 

—i +i 

(2) -i and +i Btrancb intersect ^^ 

-i +i 

(3) a strand connects —i and +i ^/^~\ 

—i +i 


Geometric interpretation of the bilinear form 18 

Set up a bijection 

p{i' + i - a", 0) ^-^p{i' - i + a", 0) (2.36) 

that takes diagrams from class (1) in the first set to diagrams of class (2) in the second 
set by adding a crossing, diagrams of class (2) to diagrams of class (1) by removing the 

„ ( ^, X_. ) . - -_ . ... ,3, . a_. . ... ,3, .. _. .e 

orientation of the +i —i strand. 

Let _Ej/+j_jj//1a = -Ej'_|_j_j1^-Ej"1a, that is, /i = A + i^ is the weight of the region to the 
right of the strand near —i 

We have in U 

-Sj'+J-iJ"lA = -£^i'-i+M"lA + [(^,/^)]j-£^i'j"lA- (2.37) 


(E,'+,_,i"lA, 1a) = (^i'-.+.i"lA, 1a) + [(2,/i)].(^i'i"lA, 1a). (2.38) 

Since deg ( y/ ^ j = deg ( \(' ^ j = 0, a diagram of class (1), respectively class 

(2), contributes to {Ei>_^_i_ii'flx, 1a)' as much as its image in class (2), respectively class (1), 
contributes to (£'j/_j_,_jj//1a, 1a)'- 

A diagrams of class (3) in p{i' + i — ii", 0) comes from some diagram in p{i'i", 0) 
by adding a {+i — i) cap to the correct position. Similarly, a diagram of class (3) in 
p{i' — i + ii" , 0) comes from a diagram in p{i'i" , 0) by adding a {—i + i) cap. Thus, the 
only contribution to the difference (i?j/_,_j_jj//lA, 1a)' — (-Ej/_j4.jj"1a, 1a)' comes from class (3) 
and is given by 

f i f^\ 


{Ei'+i-ii"lx, l\)'-{Ei'_i_^_ii"lx, 1a)' = -, _ 2 {Ei'i"l\, 1a)' -, _ 2 {Ei'i"^x, 1a)' 

^ Hi ^ Hi 


Qi ' - Qi 

(-Ej'+i_ji"lA, 1a)' = (-Ej'-j+ii"lA, 1a)' + -, i {Ei'i"l\, 1a)' 


= {Ei' -i+ii"l\., 1a)' + — Ti — (-£^i'i"lA, 1a)' 


where the coefficient of (i?j'j/'lA, 1a)' in the last term is easily recognized as [(i,/i)]j, com- 
pleting the proof of the lemma. D 


Geometric interpretation of the bilinear form 


We can finish the proof that 

(EJa, 1a)' = (^Ja, 1; 


by induction on || i ||, the length of i. During the induction step we move all negative 
entries of i to the left of all positive entries, sometimes adding terms {Ejlx, 1a)', respectively 
(ii^jlA,lA) with II j 11 = 11 i II —2 to the equation. We can then reduce to the case when 
all negative entries of i precede all positive entries, which is Lemma 12.101 Theorem 12.71 
follows. n 

We can turn this proof around to define U in a more geometric way than in Lusztig [23] . 
Start with the (Q(g)-algebra 'U with mutually-orthogonal idempotents 1a and basis {Eilx}i^\, 
over all finite signed sequences i G SSeq and A G X. The multiplication is 


Ei'ilx if/i = A + ix, 



When i is the empty sequence, -Eg 1a = 1a- 

Define a (Q(g)-bilinear form (, )' on 'U via the sum over diagrams, formula (12.241) . Let 
X C 'U be the kernel of this bilinear form. Then 

X — ^ ^Xa, 



where ^^'Ua is spanned by -EjIa for all i such that fi = X + ix- It follows from the definition 
of the bilinear form that X is an ideal of 'U. 
It is not hard to check that 

U = 'U/X and 


following the above proof of Theorem 12. 7[ This definition is not too far off from Lusztig's 
original definition, which utilizes f = U^, defined as the quotient of the free associative 
algebra 'f by the kernel of a bilinear form on 'f . The latter bilinear form is the restric- 
tion of (, )' on 'U to 'f C 'U. Here we map 'f -^ 'U by sending 6i = di^ . . . 9i^ to 
Eilx = E^i^ . . . E^i^lx for each positive sequence i (this map is not a homomorphism) , 
alternatively we can send 6i to E_ilx- 


vertical arrows \ 

do not respect I 

^algebra structure/ 

Z[g,g ^]-algebras 



For each A G X there are two inclusions (corresponding to U^ and U ) of the lower half of 
the diagram to the upper half. Restricting to weight spaces, we get the following diagram 



Ua <- 

X±ux ^ A 

-^ f,. ^ 


^ aI 


Q(g)-vector spaces 

Z[g, g ^] -modules 

where x±y^1Jx = 1a±i.xU1a, etc. 

3 Graphical calculus for U categorification 

3.1 The 2-category U 

3.1.1 Definition 

We define a 2-category U for any root datum (Y,X, {,),.. .) of type (/, •). This 2-category 
has the structure of an additive k-hnear 2-category, see [12] and [211 Section 5]. Thus the 
hom sets between any two objects form a k-linear category, and composition and identities 
are given by additive k-linear functors. The 2-morphisms in U are represented graphically 
using string diagrams, see [2T| Section 4] and the references therein. 

Definition 3.1. Let (Y, X, (,),...) be a root datum of type (/, ■). U is an additive k-linear 
2-category. The 2-category U consists of 

• objects: A for A G X. 

The homs U{X, A') between two objects A, A' are additive k-linear categories consisting of: 

• objectqj of W(A, A'): a 1-morphism in U from A to A' is a formal finite direct sum of 

Silx{t} = lySilx{t} 

for any t G Z and signed sequence i G SSeq such that X' = X + ix- 

• morphismsofW(A, A'): for 1-morphisms £^jlA{t},£^jlA{t'} &U, hom. sets U{Sil\{t},Sjlx{t'}) 
of U{X, A') are graded k-vector spaces given by linear combinations of degree t — t' 
diagrams, modulo certain relations, built from composites of: 

^We refer to objects of the category U{\, A') as 1-morphisms of U. Likewise, the morphisms of U{X, A') 
are called 2-niorphisms in U. 


The 2-category U 


i) Degree zero identity 2-morphisms Ix for each 1-niorphism x in U\ the identity 
2-morphisms 1e+iX\{t} ^^^1 '\-£_i\^{t}, for i G /, are represented graphically by 



\ + ix 






and more generally, for a signed sequence i = 61116212 ■■■ ^mim, the identity 
^£iix{t} 2-morphism is represented as 

H «2 

\ + i 


H «2 ^m 

where the strand labelled ia is oriented up ii 6^ = + and oriented down if 
6a = —■ We will often place labels with no sign on the side of a strand and 
omit the labels at the top and bottom. The signs can be recovered from the 
orientations on the strands as explained in Section 12.21 

ii) For each A G X the 2-morphisms 


1 1 
T i,x 


X i,j,X 

X( i,j,X 


A + «x- 


. A 



A - 

'A + zx 



i^ Vi 


i ■ i 

i ■ i 



















The 2-category U 


with c±i^\ defined in (12.191) . such that the following identities hold. 
The s[2 relationqj ( all of the strands are labelled by i): 

i) l\^i^£^il\ and l\£^il\^i^ are biadjoint, up to grading shifts: 

X + ix 

\ + ix 





A + «x 

\ + ix 

\ + ix 
\ + ix 

\ + ix 



\ + ix 



A \ + ix 

A \J 


\y X + ix 

iii) All dotted bubbles of negative degree are zero. That is, 
A A 

= ifa<(i,A)-l / \=0 ifa<-(z,A)-l (3.4) 


for all a G Z+, where a dot carrying a label a denotes the a-fold iterated vertical 
composite of f or i depending on the orientation. A dotted bubble of 

degree zero equals 1: 

A A 

\y =1 for(i,A)>l, \y =1 for(z,A)<-l. 



iv) For the following relations we employ the convention that all summations are 

increasing, so that X]f=o ^^ ^^^° if a < 0. 


^ The vertical ii-crossing was represented by the diagram 
crossing for simpHcity. 



-(*,A> ' 

• " 





^.. \ 


T \J 


9=0 ^^— • 




in [5T1 [22] . Here we use a standard 


The 2-category U 


(i,A>-l / 



/=0 9=0 -(«,A>-l+5 

-(^A)-l / 





/=0 9=0 («.A>-l+9 




A the dotted bubbles appearing 

for all A G X. Notice that for some values o 
above have negative labels. A composite of < 

number of times does not make sense. These dotted bubbles with negative labels 
called fake bubbles, are formal symbols inductively defined by the equation 


with itself a negative 

-(k^)-1 -(iiAf-l+l -(i,A7-l+a / \ («Va)-1 {i,X)-l+a 


and the additional condition 

A A 

1 if {i, A) = 0. 


Although the labels are negative for fake bubbles, one can check that the overall 
degree of each fake bubble is still positive, so that these fake bubbles do not 
violate the positivity of dotted bubble axiom. The above equation, called the 
infinite Grassmannian relation, remains valid even in high degree when most of 
the bubbles involved are not fake bubbles. See [21] for more details. 

v) NilHecke relations: 



if i« if >* if >i if '* 



We will also include 03. 101) for i = j as an s[2-i'elation. 


The 2-category U 


All 2-inorphisnis are cycliqj with respect to the above biadjoint structure. This is 
ensured by the relations (13.31) . and the relations 


The cyclic condition on 2-niorphisnis expressed by ( 13. 3p and ( ]3.10p ensures that 
diagrams related by isotopy represent the same 2-morphism in U. 

It will be convenient to introduce degree zero 2-morphisms: 




where the second equality in (13. lip and (13.121) follow from (I3.10p . 
For i j^ j 





The i?(i/)-relations: 
i) For i ^ j 





^ iii-j^ 0. 

X^ - X/ X/ - X 






"'See |21) and the references therein for the definition of a cychc 2-niorphisni with respect to a biadjoint 


The 2-category U 


ii) Unless i = k and i ■ j ^ 


For i ■ j j^ 




For example, for any shift t there are 2-niorphisnis 



, , : lA{t} => S-i+ilx{t - C+i,x} /~\ • ^~i+dx{t} =^ lx{t - C^i,x} 

i A 

in U, and the diagrammatic relation 

gives rise to relations in U^Smlxit}, Smlxit + 3i ■ i}) for all t G Z. 

the additive k-linear composition functor U{X, A') x U{X', A") -^ W(A, A") is given on 
1-morphisms of U by 

£jlx'{t'} X £ilx{t} ^ £jilx{t + t'} 
for ix = A — A', and on 2-morphisms of U by juxtaposition of diagrams 






The 2-category U 


Remark 3.2. By choosing an orientation of the graph associated to Cartan datum (J, ■) 
the i?(i/)-relations above can be modified by replacing them with signed -R(z^) relations 
determined by invertible elements Tjj, Tji chosen for each edge of the graph (see [T?j for 
more details). 

lA has graded 2-homs defined by 

YLOMu{x,y) := @YLomu{x{t},y). 


Also define graded endomorphisms 

END2^(x) :=HOMw(a;,x). 



The 2-category with the same objects and 1-morphisms as U and 2-homs given by H0M2^(x, y) 
is denoted W*, so that 

W{x,y) = 'iiOMu{x,y). (3.21) 

U* is a graded additive k-linear 2-category with translation PH Section 5.1]. 

3.1.2 Relations in U 

We now collect some other relations that follow from those above. These relations simplify 
computations in the graphical calculus and can be used to reduce complex diagrams into 
simpler ones (see the proofs of Proposition 13.61 and Lemma [3. 9j) . 

Proposition 3.3 (Bubble Slides). The following identities hold in U 


E(« + i-/) 



iii= j 




A + Jx 




iii-j = -l 


iU-j = 


The 2-category U 



E (a + 1 - /) 


iii = j 






if i- j = -1 


if i ■ j = 


Proof. When i = j the proof appears in [21]. For i 7^ j the equation follows from decom- 





1 f -(i,A+jx)-2+Q _(j,A+ix>-2+(J'-? 3 ^ (i,A+ix>-l+" {i,A+ix>-l+a^ ■? 

using the relations fl3.13p and (13.141) . D 

Proposition 3.4 (More bubble slides). The following identities hold in U 

( A + ix 



X + ix 


X + ix 


X + Jx 




((i,A> + l) + Q 


iii = j 

iU-j = -l 

X + jx 



{-(i,A>-l) + (a-2) 




if z = j 




/ A 

if i- j = -1 



Spanning sets of HOMs in U 


Proof. These equations follow from the previous Proposition. 
Proposition 3.5. Unless i = k = j we have 



and when i = j = k we have 

M. ^^ 

i i^ \ i --S T 



+ i: 

^ A 



where the first sum is over all fi, f2, fs, Ia > with fi + f2 + fs + fi = {h ^) and the 
second sum is over all gi,g2,g3,g4 > with gi + g2 + 93 + 94 = {i, ^) — 2. Recall that all 
summations in this paper are increasing, so that the first summation is zero if {i, A) < 
and the second is zero when (i, A) < 2. 

Reidemeister 3 like relations for all other orientations are determined from (I3.16p . (13.171) . 
and the above relations using duality. 

Proof. For (13.221) ii i y^ j post-compose both sides with the isomorphism 

X I 

j k i 


then use that i = j ^ k to apply (13.161) on the right term and establish the equality. If 
i = j then we may assume that j ^ k. In this case we pre-compose with the isomorphism 

I /V , 

i k j 

then use (13.161) on the left side to establish the identity. 
The case i = j = k appears in ^T\ Section 5.4]. 


3.2 Spanning sets of HOMs in U 

Given two Laurent power series f{q) = J2k=a fkO.^ and h{q) 
say that /(g) < h{q) if fk < hk for all k. 

J2k=a ^ith fk, hk E Z we 

3.2.1 Endomorphisms of 1^ 

For any root datum and A G X define a graded commutative ring IIa freely generated by 


.A .A 

^Qr for (i, A) > and Q^ for (i, A) < (3.24) 

{i,X)~l+a -{i,X)-l+a 

of degree ai ■ i over alH G / and a > 0. 


Spanning sets of HOMs in U 29 

Proposition 3.6. Interpreting these generators of IIa as elements of YiOViui^X) 1a) induces 
a surjective graded k-algebra homomorphism 

li^^YLOMuilxAx). (3.25) 

Proof. By induction on the number of crossings of a closed diagram D representing an 
endomorphism of 1a one can reduce D to a linear combination of crossingless diagrams 
following the methods of [211 Section 8]. Crossingless diagrams that contain nested bubbles 
can be written as linear combinations of crossingless nonnested diagrams using the bubble 
slide equations in Propositions 13.31 and 13.41 Using the Grassmannian relations (13. 7p all 
dotted bubbles with the same label i can be made to have the same orientation given by 
(1321. □ 

The image of the monomial basis of IIa under surjective homomorphism fl3.25p is a 
homogeneous spanning set of the graded vector space H0M;^(1a, 1a). Denote this spanning 
set by 50,0, A. 


oo ^ 

"■ = 11111— 2^- P'^"' 

jg/ a=l ^« 

■ji depends only on the root datum (on the values oi i ■ i over all i E I). The graded 
dimension of IIa is vr. 

Corollary 3.7. gdim HOMiY(lA, 1a) < tt and H0Mz^(1a, 1a) is a local graded ring. 

Remark 3.8. IIa is not Noetherian. 

We call monomials in this basis of IIa and their images in H0M2^(1a, 1a) bubble mono- 

3.2.2 Horns between Silx and Sjlx for positive i and j 

Recall rings i?(i^), for z/ G N[J], from [16], [17] and decomposition 

^(^)= jRi^)i, (3.27) 


where jR{v)i is spanned by braid-like diagrams D with i, j being the lower and upper 
sequences of D. Adding upward orientations to a diagram D in -R(z/), placing it to the left 
of a collection of bubbles representing a monomial in IIa 


(J,A)-l+ai -(i,A>-l+a3 4- J 


(i,A)-l + Q2 -(j,A)-l + Q4 


Spanning sets of HOMs in U 30 

and viewing the result as a 2-niorphism from £il\ to £j'^\ induces a grading-preserving 
k-linear map 

V7i,,-,A : jR{y)i ®k Ha — ^ -ROMuiSilx, SjU). (3.28) 

Lemma 3.9. v^ij,A is surjective. 

Proof. Start with a diagram D that represents an element in H0M^(£^j1a,£^j1a)- The 
relations in our graphical calculus allow us to inductively simplify Z), by reducing the 
number of crossings if a strand or a circle of D has a self-intersection or if D contains two 
strands that intersect more than once. Bubble sliding rules allow moving bubbles to the 
far right of the diagram. Eventually, D reduces to a linear combination of diagrams which 
are products of diagrams representing elements oi jR{v)i and monomials in IIa. D 

The elements in H0M^(£^j1a,^j1a) given by diagrams without circles, with no two 
strands intersecting more than once, and with all dots at the bottom are precisely the 
image under ^ij^\ of the basis jBi of j-R(z/)j described in [16]. Denote by -Bjj,a the image 
under v^j^-.a of the product basis jBi x {monomials in IIa} in jR{^)i ^ ^\- The lemma 
implies that Bijx is a spanning set in ][iOyiu{£il\, £jl\) . 


£,lx:= ^flA, £-ulx:= £-ilx. (3.29) 

«£Scq{v) JGScq(i/) 

£ul\ and £-yl\ are 1-morphisms in U. Summing V5ij,A over all i, j G Seq(z/), we obtain a 

^uX- Ri.^) ®k Ha ^ END^(^aA) = Endi,.(^,lA). (3.30) 

Proposition 3.10. Homomorphism (y9,^A of graded k-algebras is surjective. 


Adding a downward orientation to a diagram D in -R(z/), multiplying it by (—1)" where 
a is the number of crossings of identically colored lines, and placing it to the left of a 
collection of bubbles representing a monomial in IIa induces a surjective homomorphism 

</p_,,A : R{y) ®k Ha -^ ENDi,(£_,lA). (3.31) 

3.2.3 Spanning sets for general i and j 

We now describe a spanning set Bij\ in HOM^^ (£^j1a,£^j1a) for any i,j,\. Recall that 
p'{i,j) denotes the set of (i, j)-pairings, and p{i,j) is a set of minimal diagrammatic 
representatives of these pairings. For each diagram D & p{i,j) choose an interval on each 
of the arcs, away from the intersections. The basis Bij^x consists of the union, over all D, 
of diagrams built out of D by putting an arbitrary number of dots on each of the intervals 
and placing any diagram representing a monomial in IIa to the right of D decorated by 
these dots. Notice that Bijx depends on extra choices, which we assume are made once 
and for all. For example, for i = {—j, +i, +j, —i) and j = (+i, — ?) the choice of minimal 


Spanning sets of HOMs in U 31 

diagrams is unique, see example following Theorem 12. 7[ The choice of intervals for the 
dots is not unique, though. Choosing these intervals as shown below 

+i —i 



-j +i +j -i -j +i +j -i 

results in the spanning set i?(__j,+j,+j _i),(+j _i),A whose elements are 

(i,X)-l+ai -(j,A>-l+a3 \_J 

(J,A)-1+Q2 -{j,\)-l+ai 
0,2 az 

r7v\ ^ ^ 

"^ ' '. ~^ bubble monomial in Il\ 

-3 +i +3 ~i 

(i,A)-l+Qi -(j,A>-l+a3 \_J 

(J,A)-1+Q2 -(i,A)-l+a4 



bubble monomial in Il\ 

-J +i +J ~i 

over all nonnegative integers ai, 02, 03 and over all diagrammatic monomials in IIa (bubble 
orientations are for the case (i. A) > 0, (j. A) < 0, (/c. A) > 0). 

Proposition 3.11. For any intermediate choices made, -Bjj.a spans the k-vector space 

Proof. Relations on 2-morphisms in U{S^lx) allow arbitrary homotopies of colored dotted 
diagrams modulo lower order terms, i.e., terms with fewer crossings, fewer circles, etc. 
Detailed discussion in [21], Section 8] generalizes to the present situation without difficulty. 


For each s G Bij\ its degree deg(s) is an integer, determined by the rules in Section ISTTl 

Proposition 3.12. For any i,j, and A 

niEilx,EjU)= J2 ^'''^'^- (3-32) 


* 5 J , A 


Spanning sets of HOMs in U 32 

Proof. The LHS of the equation (13.321) equals vr times the RHS of the formula (12.221) . This 
vr is matched in the RHS of (13.321) by the summation over all monomials in 11;^, since q 
to the degree of these monomials add up to tt. The product term in the RHS of (12.22^ is 
matched by the contribution to the RHS of (13.321) by all possible placements of dots. For 
each i-labelled strand dots contribute 



1 - q" 1 - ii 

to the product, since the degree of a dot is z ■ i, and the sum is over all ways to put some 
number a of dots on this strand. Finally, the sums over all minimal diagrams D G p{i,j) 
give equal contributions to the two sides of (I3.32p . D 

Remark 3.13. In view of the first equality in (12.221) . we can restate the above proposition 
via (, ) in place of (, ) on the LHS. 

Notice that 

gdim,(HOM^,(£,l,,£:,-l,)) = J2 ^"''^'^ (3-33) 

if and only if Bij\ is a basis of B.OMu {Silx, Sjlx)- 
Corollary 3.14. For any sequences i,j and A G X 

gdimk(HOM,,(^aA,^,-lA)) < n{Eilx,Ejlx). (3.34) 

Definition 3.15. We say that our graphical calculus is nondegenerate for a given root 
system and field k if for all i,j, A the set -Bjj,a is a basis of HOM^^ (^jIa, ^jIa)- 

Thus, a calculus is nondegenerate if the equality holds in Corollary 13. 141 for all i,j,X. 

Remark 3.16. Nondegeneracy holds if the above condition is true for all A G X and all 
pairs of positive sequences i,j. 

3.2.4 Endomorphisms of £y _y'lA 

For z/,i/' G N[J] let 

^,,_,,1a:= Si^jlx. (3.35) 


Consider the graded ring E'NDu^S^^^'l x) ■ A spanning set for this ring is given by dotted 
minimal diagrams of {i{—i'),j{—j'))-pa.iYmgs over all i,j G Seq(i/), i',j' G Seq(z/') times 


Spanning sets of HOMs in hi 


bubble monomials in IIa. 


(J,A>-l+ai -0',A>-l+a3 4- J" 


(J,A)-l+a2 ~{j,\)-l+a4. 

Let Xu^-ui^x be the subspace spanned by diagrams which contain a U-turn, i.e. an arc with 
both endpoints on R x {1} or on R x {0}. 

Proposition 3.17. Xy^_yi\ is a 2-sided homogeneous ideal of END^(£^,^ _,^/1a) which does 
not depend on choices of minimal diagrams for pairings. 

Proof. Left to the reader. 


Denote by R^^^^'^x '■= ^^^u{^u~u'^x)/^u,~u',\ the graded quotient ring, and by /3 the 
quotient map. There is a homomorphism 

a : R{iy) ®k i?(z/') ®k Ux -^ END^ (£:,,_,, 1a) (3.36) 

given by placing diagrams representing elements in R{v), R{i^'), Ha in parallel 



(i,A)-l+Ql -(j,A)-l+Q3 4- J 


X -l X- («,A)-1 

o a 

+02 -(j,A)-l+a4 

orienting diagrams for R{i') upwards and diagrams for -R(z^') downwards, and multiplying 
by (—1)", where a is the number of crossings in D' of equally labelled strands. The 
composition f3a of this homomorphism with the quotient map is surjective. The diagram 
below contains an exact sequence of a ring, its 2-sided ideal, and the quotient ring: 

R{u) ® R{u') ® Ha 




¥ Ru,-u'x — > . 



Properties and symmetries of 2-category hi 34 

Remark 3.18. If the graphical calculus is nondegenerate, j3a is an isomorphism, and the 
sequence splits 

y X.,-.',A y YMJ^uiSu^-u'U) ^ RW) ® R{iy') ® Ha y O. (3.38) 

A split ring homomorphism induces a split exact sequence of Grothendieck groups [28l 
Section 1.5] 

>KoiIu,-u',x) > Ko{ENDu{£u,-u'lx)) < > Ko{R{u) R{u') ^U^) ^0 , 


leading to a canonical decomposition of the middle term as the sum of its two neighbors. 

3.3 Properties and symmetries of 2-category U 

3.3.1 Almost biadjoints 

The 1-morphism S+ilx does not have a simultaneous left and right adjoint S-ilx-^-i^ be- 
cause the units and counits which realize these biadjoints in U* are not degree preserv- 
ing. However, if we shift £-ilx^i^ by {— c+j^a}, then the unit and counit for the ad- 
junction S-^-ilx H S^ilx+ix{~c+i^x} become degree preserving. More generally, we have 
S-^-ilx{t} H S-ilx+ix{~c+i^x — t} inU since the units and counits have degree: 

\^ 1 = c+,,A + (-c+.,a) = (3.40) 

i A / 

deg ^r> = (c-i,A+ix)) - (-c+.,a) = (3.41) 

\ +i -i {t-c+i^x-t}/ 

and still satisfy the zigzag identities. Similarly, £^ilx{t} has a left adjoint S_ilx+ix{c+i,\ — 
t} in U. One can check that with these shifts the units and counits of the adjunction 
S_ilx+ix{c+i,x ~ t} ~^ ^+i^x{t} become degree zero and are compatible with the zigzag 

The left adjoint S^ilx+ix{c+i,x — t} and the right adjoint S^ilx+ix{~c+i^x — t} of 
S^ilx{t} only differ by a shift. We call morphisms with this property almost biadjoint. 
This situation is familiar to those studying derived categories of coherent sheaves on Calabi- 
Yau manifolds. Functors with these properties are called 'almost Frobenius functors' in 
[Is] where several other examples of this phenomenon are also given. 

It is then clear that S^ilx{t} and S-ilx{t} have almost biadjoints in U for all t G Z 
and A G X with 

Ia^-jIa+jxI'^+J.A ~ ^ ^X+ix^+i^x{t} H ^X^-i^X+ix{~'^+i,X —t} C^AO) 

lA^+jlA-ixI'^-i.A ^ ^ ^X-ix^-i^x{t} ~^ ^xS+i^X-ix{~(^-i,X — t} ■ 


Properties and symmetries of 2-category U 


Every l-morphisin in W is a direct sum of composites of £+il\{tYs and £-il\{tYs together 
with identities; by composing adjunctions as explained in [211 Section 5.5], the right and 
left adjoints of £il\{t} can be computed. Thus, it is clear that all 1-morphisms in U have 
almost biadjoints. 

Positivity of bubbles: W(1a, 1a{0) = if t > 0, andW(lA, 1a) is at most 1-dimensional 
(isomorphic to k if the calculus is nondegenerate). 

3.3.2 Symmetries of W 

We denote by U°^ the 2-category with the same objects as lA but the 1-morphisms re- 
versed. The direction of the 2-morphisms remain fixed. The 2-category IA^° has the same 
objects and 1-morphism as U, but the directions of the 2-morphisms is reversed. That is, 
U^°{x,y) = U{y,x) for 1-morphisms x and y. Finally, U™°p denotes the 2-category with 
the same objects as U, but the directions of the 1-morphisms and 2-morphisms have been 

Using the symmetries of the diagrammatic relations imposed on U we construct 2- 
functors on the various versions of U. In Proposition 13.281 we relate these 2-functors to 
various Z[g, g~^]-(anti)linear (anti) automorphisms of the algebra U. The various forms of 
contravariant behaviour for 2-functors on U translate into properties of the corresponding 
homomorphism in U as the following table summarizes: 


Algebra maps 


Z[g,g~^]-linear homomorphisms 


Z g, q^^ -linear antihomomorphisms 


Z[g, g~^]-antilinear homomorphisms 

u -^ w=°°p 

Z[g, g~^]-antilinear antihomomorphisms 

Rescale, invert the orientation, and send A ^-* —A: Consider the operation on the 
diagrammatic calculus that rescales the u-crossing ^>\ ■ ■ \^^ ~ /\ • • a ^°^ alH G / and 
A G X, inverts the orientation of each strand and sends A ^^ —A: 

/ +i +A:+fe -3 -k -j 


\ +i +k +, 

-j -k -k +j +k +j 

] -■!■ -J -] 

-i -k -j +i +j +j 


Properties and symmetries of 2-category U 


This transformation preserves the degree of a diagram, so by extending to sums of diagrams 
we get a 2-functor uj:U-^U given by 





It is straight forward to check that cj is a strict 2-functor. In fact, it is a 2-isomorphism 
since its square is the identity. 

Rescale, reflect across the y-axis, and send A 


The operation on diagrams 

that rescales the ii-crossing ^x^ ax^^ ~ /^ a x^*^"^ ^^ ^ ^ ^ ^"^^ A G X, reflects a diagram 
across the y-axis, and sends A to —A leaves invariant the relations on the 2-morphisms of 
U. This operation 


\ +i +k +j -i -j -j 


^j -J -i +3 +k +i 

is contravariant for composition of 1-morphisms, covariant for composition of 2-morphisms, 
and preserves the degree of a diagram. Hence, this symmetry gives a 2-isomorphism 

A ^ -A 

l^^si^S2 ■ ■ ■^Sm-i^SmlAl^l ^^ 1-A^Sm^; 

Sm — 1 

■ Cs2^si -L— ^i'-/ 

and on 2-morphisms a maps linear combinations of diagrams to the linear combination 
of the diagrams obtained by applying the above transformation to each summand. The 
relations on U are symmetric under this transformation, and ? is a 2-functor. The square 
of a is the identity. 

Reflect across the x-axis and invert orientation: Here we are careful to keep track 
of what happens to the shifts of sources and targets 


/ +i +k+k -j -k -3{t'} \ 
\ +i +k +j -i ~j -j {t} ) 

+i +fc +i -i -j -3{-t] 

+j +k +k -j -k -j{-t'} 


Properties and symmetries of 2-category U 


The degree shifts on the right hand side are required for this transformation to preserve 
the degree of a diagram. This transformation preserves the order of composition of 1- 
morphisms, but is contravariant with respect to composition of 2-morphisms. Hence, by 
extending this transformation to sums of diagrams we get a 2-isomorphism given by 

A ^ A 


and on 2-morphisms ip reflects the diagrams representing summands across the x-axis and 
inverts the orientation. Again, the relations on U possess this symmetry so it is not difficult 
to check that ■?/' is a 2-functor. Furthermore, it is clear that if) is invertible since its square 
is the identity. 

It is easy to see that these 2- functors commute with each other 'on-the-nose'. That is, 
we have equalities _ _ _ _ 

uia = ua, aip = ipa, ujip = ip^ (3.45) 

The composite 2-functor ijjua is given by 

A ^ A 

l^£si£s2-- ■£s^.i£s^'i-x{t} ^ lx£-s^£- 

^771 — 1 

' ^—S2^-Si^fJ,\ tj 

and is given on 2-morphisms by rotating diagrams by 180°. 

The following transformation only differs from ipua by a shift and is given by taking 

Rotation by 180° (taking right adjoints): This transformation is a bit more subtle 
because it uses the almost biadjoint structure oiU, in particular, the calculus of mates (see 
PT| Section 4.3]). For each 1^x1 a G U denote its right adjoint by Ixyl^. The symmetry 
of rotation by 180° can also be realized by the 2-functor that sends a 1-morphism I^sjIa 

to its right adjoint IxV^fi and each 2-morphism (: 
adjunctions lf,xlx H IaZ/I^ and If^x'lx H Ixy'lf,- 
(*. Pictorially, 






lfj,x 1a to its mate under the 

That is, ( is mapped to its right dual 






Karoubi envelope, lA, and 2-representations 38 

This transformation is contravariant with respect to composition of 1-morphisms and 2- 
morphisms. We get a 2-functor 

A ^ A 

C ^ C* (3.46) 

where the degree shift t' for the right adjoint IxS^sm^-s^-i ' ' ^^~s2^-sl^^i{—t + t"}, deter- 
mined by fl3.42p . ensures that r is degree preserving. Inspection of the relations for U will 
reveal that they are invariant under this transformation so that r is a 2-functor. 

We can define an inverse for r given by taking left adjoints. We record this 2-morphism 

A ^ A 

C ^ X- (3.47) 

with degree shift t" determined from f l3.42p and the left dual *( of the 2-morphism ( defined 
in [211 Section 4.3]. 

Remark 3.19. The composition Ti/juja: U ^U gives 2-isomorphism that fixes all diagrams 
and only effects the grading shifts. 

Remark 3.20. There are degree zero isomorphisms of graded k- vector spaces 

U\fx,y) ^ W*(x,r(/)i/), (3.48) 

W{x,gy) ^ W{T~\g)x,y), (3.49) 

for all 1-morphisms /, g, x, y in U*, defined at the end of Section [3.1.11 

3.4 Karoubi envelope, U, and 2-representations 

The Karoubi envelope Kar{C) of a category C is an enlargement of the category C in which 
all idempotents split (see [211 Section 9] and references therein). There is a fully faithful 
functor C -^ Kar{C) that is universal with respect to functors which split idempotents in 
C. This means that if F: C — > P is any functor where all idempotents split in D, then F 
extends uniquely (up to isomorphism) to a functor F: Kar{C) —>■ T) (see for example [4], 
Proposition 6.5.9). Furthermore, for any functor G: C ^ V and a natural transformation 
a: F ^ G, a extends uniquely to a natural transformation 5 : F => G. When C is additive 
the inclusion C -^ Kar{C) is an additive functor. 


Karoubi envelope, U, and 2-representations 


Definition 3.21. Define the additive k-linear 2-category U to liave tlie same objects as 
U and liom additive k-linear categories given by U{X,X') = Kar {U{X,X')). The fully- 
faithful additive k-linear functors U{X, X') -^ W(A, A') combine to form an additive k-linear 
2-functor U ^ lA universal with respect to splitting idempotents in the hom categories 
W(A, A'). The composition functor U{X, A') x W(A', A") -^ W(A, A") is induced by the uni- 
versal property of the Karoubi envelope from the composition functor for U. The 2-category 
U has graded 2-homs given by 





Definition 3.22. A 2-representation of W* is a (weak) graded additive k-linear 2-functor 
^* : U* -^ M*, where A^* is a graded additive k-linear 2-category with a translation. 

A 2-representation of U is an additive k-linear 2-functor '^ : lA ^ M. that respects the 
grading. This happens when there is an additive k-linear 2-functor M. -^ M.* ^ with M.* a 
graded additive k-linear 2-category, making the diagram 






weakly commutative. Thus, to study 2-representation oiU it suffices to study 2-representation 
of U* and then restrict to degree-preserving 2-morphisms. 

A 2-representation of W is a additive k-linear 2-functor "$ : U ^ Ai that respects the 

Let A^ be a 2-category as above in which idempotents split. Any 2-representation 
"$*: U* -^ Ai* gives a unique (up to isomorphism) 2-representation \1/: W ^ Ai. The 
2-functor ^ is obtained from \1'* by restricting to the degree preserving 2-morphisms of W* 
and using the universal property of the Karoubi envelope. This is illustrated schematically 



restrict to 



Karoubian envelope 

Remark 3.23. The 2-functors u, a, ip, r on U extend to 2-functors on U, for which we 


Direct sum decompositions 40 

use the same notations. For example, 

u:U -^ U 
A ^ -A 

C ^ ^(C) (3.52) 

and the other 2-morphisms a, ip, and r are defined analogously. In particular, each 1- 
morphism in U has left and right adjoints. 

3.5 Direct sum decompositions 

Recall that £u^\ is the direct sum of £^j1a over all i G Seq(z/): 

£,lx:= Silx- (3.53) 

Due to the existence of the homomorphism ip^^x in the formula (13.301) any degree idem- 
potent e of R{i^) gives rise to the idempotent y:>ij,x{e) of Sylx and to the 1-morphism 
{£^lx,'fu,x{e)) oiU. 

Introduce idempotents 

similar to the idempotent Cj^m in [HI Section 2.2] and [T7]. Define 1-morphisms £^^_j(m) 1a 
and 8_i{m)lx in U by 

t+j(m)lA := [t+im.lx, e+i^rn) i — } , (3.54) 

( 171(1 — m)i ■ il 
£_i(m)lx ■= {£-imlx,e-i^rn) < ^ ^ > ■ (3.55) 

As in [ini [I7| , we have direct sum decompositions 

For any divided power sequence i = (eii^ ,6212 \ ■ ■ ■ i^m^m ) define 

^jIa := (^••lA,ei), 


Direct sum decompositions 41 

where i is the sequence 

{eih, ..., 6111,6212, ■ ■ ■ 6212, ■■■, 6mim ■■■, 6mim) = {{^IHT^ (£2^2)"' • • • (f^m^m)"™) , 

with term 6iii repeating ai times, term £2^2 repeating 02 times, etc., and 

is the horizontal product of idempotents. 

When interested in only one part of a sequence i, we write . . .i" . . . instead oi i = 
% i" %" and E i" \\ instead oi Ei\\ = Si'i"i"'lx- 

Proposition 3.24. For each i,j & I, i ^ j, and A G X there are 2-isomorphisms of 
1-morphisms in U 

(37 ^...+j(2")+j + i('«+l-2a) 1a — (^ O ^_j(2a + l)_|_j_|_j(d-2a) 1a, 

a=0 a=0 

L^J LiJ 

\JJ ^...-j(2a)_j_j(d+l-2a) 1a — ("T^ C^..._i(2o+l) _ j_ j(d-2o) 1 A , (3.57) 

a=0 a=0 

where d = dij = —{i,jx)- 

Proof. These isomorphisms follow from categorified quantum Serre relations p^ Proposi- 
tion 2.13] and [T71 Proposition 6] between idempotents in rings -R(i^), via homomorphisms 

ip^^x and V2-,.,A- □ 

Proposition 3.25. For each i & I, X ^ X there are 2-isomorphisms in U 

Si'+i_ii"lx = Si'_i^ii"lx ®[{i,,j.)]i ^i'i"'^\ if {h A + i"x) > 

Si'_i+ii"lx = Si'+i_ii"lx ®-[(i,fj.)], Si'i"lx if {i, A + i"x) < 

where /x = A + i'x- 

Proof. Set X + i"x = ^i. The decomposition Si'+i^a^lx = Si'-i+w'lx ©[(i,/.)], ^i'j"lA for 
(i, /i) > is given by 2-morphisms in U 

a : ^i'+i-n"lA ^ Si'_,^ii"lx ©[(i,M>]i ^j'i"lA (3.58) 

a~ : £i'-i+ii"'^x®[{i,ti)U £i'i"'^x ^ £i'+i-ii"'^x (3.59) 

where a and a^^ consist of matrices of diagrams 




a, := ^ -(ui-^+J for < s < {i, fi) - 1, 



Kq{IA) and homomorphism 7 42 

Note that all bubbles that appear in as above are fake bubbles. One can check that 
a~^a = Idr., ...I, and that aa~^ = Idg., .„i>ffi,,- >,i, using the sto-relations (for details 
see [2lj)- Here we have taken 

©[(i,^)] 1a = 1a{1 - («, /i)} © ■ ■ ■ © lx{2s + 1 - (2, /i)} © ■ ■ ■ lx{{i, fi) - 1} (3.60) 

so that a and a^^ have degree zero. The isomorphism £^j'_j+jj"1a — <£^j'+j_jj"1a ©-[(i,^)]^ 
Si'i"lx for {i, X + i'x) < is given similarly (see [21]). D 

Proposition 3.26. For each i,j & I, i ^ j, X & X there are 2-isomorphisms 

S...+i.j..lx = S...-j+i..lx (3.61) 

Proof. The degree zero 2-isomorphism S,,,+i-j,,l\ = S,,,^j-^-i,,lx for i ^ j is given by maps 

i/%^i ■ £...+i-j...'^x -^ S,„^j+i„lx 
.'%<' A : £,,_j+j...1a — > £^...+j_j...1a 

To see that these maps are isomorphisms use (I3.13p . D 

3.6 Kq{U) and homomorphism 7 

Kq{U) can be viewed as a pre-additive category or, alternatively, as an idempotented ring. 
When thought of as a category, it has objects A, over all A G X. The abelian group of 
morphisms Ko(U{X, //)) is defined as the (split) Grothendieck group of the additive category 
U{X,fi). The spht Grothendieck group Ko{A) of an additive category A has generators 
[P], over all objects P of A, and relations [P] = [P'] + [P"] whenever P = P' ® P". In the 
case of U{X, n) the generators are [£ilx{t}, e], where /x = A + ix, t G Z, and 

e e End^(faA{i}) = End^(^aA) = End^,(faA) (3.62) 

is an idempotent (degree zero idempotent when viewed as an element of the larger ring 
ENDiY(^jlA))- The defining relations are 

[£ilx{t}, e] = [£Ax'{t'}, e'] + [£inlx"{t"}, e"] (3.63) 

whenever there is an isomorphism in U{X,fi) 

i£ilx{t}, e) ^ {£i'lx'{t'}, e') © (^^-Ia-I^"}, e")- (3.64) 

Moreover, Kq(U{X, fi)^ is a Z[g, g~^]-module, with multiplication by q coming from the 
grading shift 

[£ilx{t + 1}, e{l}] = q[£ilx{t}, e]. (3.65) 

We write [^^jIa] instead of [^^jIa, 1], where 1 is the identity 2-morphism of ^^jIa. 


Kq{U) and homomorphism 7 43 

The space of horns between any two objects in U{X,fi) is a finite-dimensional k-vector 
space. In particular, the Krull-Schmidt decomposition theorem holds, and an indecompos- 
able object oiU{X,fi) has the form {Silx{t},e) for some minimal/primitive idempotent 
e. Any presentation of 1 = ei + ■ ■ ■ + Ck into the sum of minimal mutually-orthogonal 
idempotents gives rise to a decomposition 


Silx{t}^^{Silx{t},er) (3.66) 


into a direct sum of indecomposable objects of W(A, /i). Any object of W(A, /i) has a unique 
presentation, up to permutation of factors and isomorphisms, as a direct sum of indecom- 
posables. Choose one representative b for each isomorphism class of indecomposables, up 
to grading shifts, and denote by B{X, fi) the set of these representatives. Then {[b]}b is a 
basis of Kq(U{\, fi)^, viewed as a free Z[g, g~^]-module. Composition bifunctors 

U{X, A') X W(A', A") — > U{\ A") (3.67) 

induce Z[g, g^^]-bilinear maps 

Kq{U{\ A')) ® Ko{Ui\', A")) -^ Ko{Ui\, A")) (3.68) 

turning Kq{U) into a Z[g, g^"'^]-linear additive category with objects A G X. Alternatively, 
we may view Kq{U) as a non-unital Z[g, g~^]-algebra 

i^o(W)(A,/i) (3.69) 

with a family of idempotents [1a]. The set B := ^^ ^^ B{\, fi) gives rise to a basis 

[B] := {[&]}fe£g of idempotented Z[g, g~^]-algebra Kq{U). Notice that basis elements are 
defined up to multiplication by powers of g; we will not try to choose a canonical grading 
normalization here. Multiplication in this basis has coefficients in N[g, g~^]. 

Both _4U and Kq(U) are idempotented Z[g, g^^]-algebras, with the idempotents 1a and 
[1a] labelled by A G X. To relate the two algebras, send 1a to [1a] and, more generally, 
Eilx to [Silx] for all i G SSeqd. 

Proposition 3.27. The assignment Eilx — > [^i'i-x] extends to Z[g, g^^]-algebra homo- 

7:^U^i^o(W). (3.70) 

Multiplication by q corresponds to the grading shift {1}. 

Proof. Kq{U) is a free Z[g, g^^]-module, so it is enough to check that the assignment above 
extends to a homomorphism of Q(g)-algebras 

7Q(g) : t^KoiU) ^nm-i] Q{q) (3.71) 


Kq{IA) and homomorphism 7 


(^U is also a free Z[g, g~^]-module, but this fact is not needed in the proof). Propo- 
sitions 13.241 I3.25[ and 13.261 show that defining relations of U lift to 2-isoniorphisms of 
1-morphisms in U and, therefore, descend to relations in the Grothendieck group Kq{U). 
Restricting 7(Q(g) to yiU gives a homomorphism of Z[g, g~^]-algebras with the image of the 
homomorphism lying in Kq{U). D 

For each X, fi & X homomorphism 7 restricts to a homomorphism of Z[g, g^^]-modules 


l^UU)lA^i^o W(A,/i) 

Proposition 3.28. Homomorphism 7 intertwines (anti) automorphisms ip, uj, a, r of ^U 
with (anti) automorphisms [ip], [uj], [a], and [r] of Ko(U), respectively, i.e., the following 
diagrams commute. 









[ip] denotes the induced action of ip on the Grothendieck group, etc. 
Proof. The proof follows from definitions and our construction of 7. 

The 2-isomorphisms u), a, ip on U* give isomorphisms of graded k-vector spaces 

U*{x,y) ^ W{u{x)My)), 

U*{x,y) = {Wr{ajx),a{y)) = U* (ajx) My)) , 

U*{x,y) = {Wr{ij{x)My)) = W{^{y),i^{x)). 
On Grothendieck rings these isomorphisms give equalities 







which should be compared with Propositions 26.1.4 and 26.1.6 in [21] and property (v) of 
the semilinear form. The last equality expressed in terms of the bilinear form (, ), with x 
replaced by ip{x), gives the identity {x,y) = {y,x). 


Idempotented rings 45 

3.7 Idempotented rings 

An idempotented ring A is an associative ring, not necessarily unital, equipped with a 
system of idempotents {Ix}, over elements x of some set Z. We require orthogonality 
Ixly = Sx,ylx and decomposition 

A = lyAl,. (3.80) 


By a (left) A-module we mean an A-module M such that 

M = 01,M. (3.81) 


In this paper three collections of idempotented rings appear: 

• Lusztig's U and its integral form _4U. Here Z = X, the weight lattice, 

U = VUIa, ^U = 1^(^U)1a. (3.82) 

• The Grothendieck groups Ko^KariU*)) and Ko(U), the latter defined in Section [321 
Again, the parameterizing set Z = X. We only study 

^o(Z^)= [l^]iro(W)[lA], (3.83) 


with {[lA]};^gj(^ being the system of idempotents in Kq{IA). Map 7: ^U — > Kq{U) is 
a homomorphism of idempotented rings. 

• For each A, /x G X the Z-graded ring 

^Ul := ^ROMu{Silx,SjW, (3.84) 

where the sum is over all i,j G SSeq with ix,jx = A — /i. Thus, the sum is 
over all sequences such that Eil\, Ejl\ have left weight ^. The parameterizing set 
Z = {i E SSeq \ ix = \ — /i}. 

The category U{X,fi) is equivalent to the category of right finitely-generated graded 
projective ^W^-modules and grading preserving homomorphisms. The equivalence functor 

W(A, /i) — > pmod-^W; (3.85) 

takes Silx to 



Surjectivity of ^ 46 

and, more generally, an object {£ilx,e) to 


The Grothendieck group Kq(1A{X, /i)) is isomorphic to the Grothendieck group of pmod— ^WjJ^ 

Notice that we get idempotented rings from the 2-category 14 in various ways: 

1) as the Grothendieck ring/pre-additive category Kq{U) of W, 

2) as rings associated to categories U{X,fi). 

The 2-category U can itself be viewed as an idempotented monoidal category. We 
encode these observations into a diagram 

(small) pre-additive idempotented additive 

2-categories monoidal categories 

Grothendieck // \\ Categories/rings of 

category/ring | | homs between objects 

(small) pre-additive categories < > idempotented rings 

3.8 Surjectivity of 7 

To prove surjectivity of 7 we will analyze the diagram of /^-groups and their homomor- 
phisms induced by the diagram fl3.37p and trace minimal idempotents there, but first recall 
some basics of Grothendieck groups of finite-dimensional algebras (as a model example) 
and graded algebras. 

3.8.1 Kq of finite-dimensional algebras 

A homomorphism of rings a : A ^ B induces a homomorphism of /^o-groups 


of finitely-generated projective modules. For definition and properties of Kq we refer the 
reader to [28], [321 Chapter II]. 

Assume that A and B are finite-dimensional k-algebras, for a field k, and a is a k- 
algebra homomorphism. If a is surjective then KQ{a) is surjective as well. On the level 
of idempotents, if 1 = ei + ■ ■ ■ + e^ is a decomposition oi 1 E A into a sum of mutually- 
orthogonal minimal idempotents, then Acs is an indecomposable projective A-module, 


Surjectivity of^ 47 

Ko{A) is a free abelian group with a basis {[Ae^jlrgs, for a subset S C {1, . . . , k}. S is 
any maximal subset with the property that Acs ^ Aet as left A-modules for any s,t & S, 
s ^ t. Applying a to the above decomposition results in the equation 

B3l=a{ei) + --- + a{ek). (3.86) 

where each a{es) is either or a minimal idempotent in B. Relabel minimal idempotents so 
that a(ei), . . . , a{em) 7^ 0, a(em+i) = ■ ■ ■ = a(efc) = (elements of S get permuted as well). 
Then 1 = a(ei) + - ■ ■ + a{em) is a decomposition of 1 G -B into a sum of mutually orthogonal 
minimal idempotents, Ba{er) is an indecomposable projective -B-module, 1 < r < m, and 

{[Ba{er)]}resn{i,...,m} (3.87) 

is a basis of Kq{B). 

Remark 3.29. If A is a finite-dimensional k-algebra, the quotient map A — > A/ J {A), 
where J{A) is the Jacobson radical of A, induces an isomorphism of KQ-gioxvps 

Kq{A)^Kq{AIJ{A)). (3.88) 

Indeed, J{A) is a nilpotent ideal, J{A)'^ = for sufficiently large A^, and the quotient by 
a nilpotent ideal induces an isomorphism of A'o's, see [32l Chapter II, Lemma 2.2]. 

Proposition 3.30. If A, B are finite-dimensional k-algebras such that all simple A- and 
i?-modules are absolutely irreducible over k, then 

KoiA)0^KoiB) = KoiA0tB) (3.89) 

via an isomorphism which takes [P] ® [Q] for projective A, respectively B, modules P and 

Qto [P(S)tQ]. 

Proof. By passing to A/J{A), B/J{B) and using the above remark, we reduce to the case 
of semisimple A and B. Then both A and B are finite products of the field k and the 
proposition follows. D 

3.8.2 Kq of graded algebras 

From here on we only consider Z-graded k-algebras, for a field k. For a Z-graded k-algebra 
A = (Bae^Aa denote by Ko{A) the Grothendieck group of finitely-generated graded left 
projective A-modules. -^"0(^4) is a Z[g, g~^]-module. 

Throughout this subsection we assume that all weight spaces Aa are finite-dimensional, 
and the grading is bounded below: Aa = for all a -C 0. 

Let PI{A) be the set of isomorphism classes of indecomposable graded projective A- 
modules, up to a grading shift. We can normalize the grading and choose one representative 
Q for each element of PI {A) so that is the lowest nontrivial degree of Q. We write 

Q e PI{A). 


Surjectivity of^ 48 

Proposition 3.31. For A as above, Ko{A) is a free Z[g, g~^]-module with tlie basis 


Proof. Since each weight space of A is finite-dimensional, the Krull-Schmidt property holds 
for graded projective finitely-generated A-modules. Any such module has a unique, up to 
isomorphism, decomposition as a direct sum of indecomposables, and Kq{A) is a free 
abelian group with a basis labelled by isomorphism classes of indecomposable projectives. 
Boundedness of A from below ensures that an indecomposable projective is not isomorphic 
to itself with a shifted grading, implying that Kq{A) is a free Z[g, g~^]-module and the rest 
of the proposition. D 

We say that a 2-sided homogeneous ideal J of A is virtually nilpotent if for any a ^TL 
the weight space {J^)a = for sufficiently large N. 

Proposition 3.32. For A as above and J a virtually nilpotent ideal of A the quotient 
map A — >A/J induces an isomorphism Ko{A) = Kq{A/ J). 

Proof. Proposition follows from the lifting idempotents property. This is the graded version 
of Lemma 2.2 in [321 Chapter 2]. D 

Proposition 3.33. Let a : A — >B be a surjective homomorphism of finite-dimensional 
graded k-algebras. Then the induced map K^la) is surjective. 

Proof. The argument is essentially the same as in the nongraded case discussed earlier. 
In the bases of Ko{A) and Ko{B) given by indecomposable projective modules, the map 
Ko^a) sends some basis elements of Ko{A) to and the rest go bijectively to the basis of 
Kq{B) (possibly after grading shifts). D 

Finally, we discuss Kq of graded idempotented algebras. Let A be an associative graded 
k-algebra, possibly nonunital, with a family of mutually-orthogonal degree idempotents 
Ix & A,x E Z, such that 

A = (^^y lyAlx 


(compare with the definition of idempotented ring in Section ISTTj) . We say that A is a graded 
idempotented k-algebra. By a graded finitely-generated projective A-module we mean a 
homogeneous direct summand of a finite direct sum (with finite multiplicities) of graded 
left A-modules Al^-jt}, over x E Z and t G Z. By Ko{A) we denote the corresponding 
Grothendieck group, which is again a Z[g, g"^]-module. 

We assume that for each x,y E Z the graded k- vector space lyAl^ is bounded below 
and has finite-dimensional weight spaces. 

Proposition 3.34. For A as above, Ko{A) is a free Z[g, g^^]-module with a basis given 
by isomorphism classes of indecomposables, up to grading shifts. 

Proof. The proof is essentially the same as that of Proposition I3.31[ The difference is in 
the absence of a canonical grading normalization for a representative Q of an isomorphism 
class of indecomposables up to grading shifts. This normalization can be chosen ad hoc, 
of course. D 


Surjectivity of ^ 49 

3.8.3 A triangle of A^o's 

We will work in the graded case, so that the rings are Z-graded and Kq are Z[g,g~^]- 
modules. Consider the diagram of Z[g, g~^]-modules 


Ko{ENDui£,,-,>lx)) ^^i?o(i?.,-.',A) (3.90) 

given by applying the Kq functor to the commutative triangle in fl3.37p . 

Recall that in [HI [17] we constructed an isomorphism of Z[g, g~^]-modules 

Ko{R{iy)) = Afu, (3.91) 

where j^^^, is the weight u summand of the 7j[q, g~^]-algebra ^if . Likewise, 

Ko{R{u'))^J,,. (3.92) 

i?(z/), respectively -R(z/'), is a free finite rank graded module over its center Z{R{v)) = 
Sym(i/), respectively Sym(z/'), isomorphic to the k-algebra of polynomials in several homo- 
geneous generators, all of positive degree. Hence, R{i^) ® R{^') C?) 11^ is a free finite rank 
graded module over the central graded polynomial algebra Sym(z/) ® Sym(z/') ® H^. This 
algebra contains a homogeneous augmentation ideal Sym"*" of codimension 1. Let 

J = [R[u) ® R[u') ® Ha) Sym+ (3.93) 

be the corresponding 2-sided ideal of R^u) ®-R(z/') ®Yi\, and consider the quotient algebra 

R ■= R{u) ® R{u') ® Ux/J = [r{u)/ {R{u) ■ Sym+(z/)) ) ® [rW)/ {R^) ■ Sym+(z/')) ). 

i? is a finite-dimensional k-algebra, and the quotient map 

a' : R{u) R{u') O Ha — > R (3.95) 

induces an isomorphism of i^o-groups 

KQ{a') : KQ{R{y) ® RW) ® Ha) -^ MR), (3.96) 

since the ideal J is virtually nilpotent, see Proposition 13.321 

We proved in [161 [I3| that any simple graded i?(z/)-module is absolutely irreducible for 
any field k, same for simple graded i?(i/)/(i?(i/)Sym"'")-modules. Also note that A'o(nA) = 
Z[g, g"^], since IIa is a graded local ring. The chain of isomorphisms 

Ko{R{u) ® R{u') ® Ha) = MR) = MR{^)) ® MRW)) = A^u ® a^w (3.97) 

establishes the following result: 


Surjectivity of ^ 


Proposition 3.35. There is a canonical isomorphism 

Kq {R{u) ® R{u') ® Ux) = Ai. ® A^u' (3.98) 

induced by isomorphisms Kq^R^u)) = js^y and Ki^{R{y'y) = js^yi constructed in p^ [T7]. 
This proposition gives us a grip on the top term in the diagram (13.901) . 

Proposition 3.36. Ki^(^l5a) is surjective. 

Proof. Since /3a is surjective, [3a{J) is a 2-sided ideal of R^^^^'^x- Start with a commutative 
square of surjective algebra homomorphisms 

R{u) ® R{u') ® H; 


-»■ Ru.-u'. 



-^ R 




and apply functor Kq to obtain a commutative diagram 




-> iro(i?.,-..,A/(/3a(J))) 

The vertical arrows are isomorphisms since J and (3a{J) are virtually nilpotent ideals. The 
bottom arrow is surjective, by Proposition 13.331 since R and Ru-u',x are finite-dimensional 
over k. Surjectivity of the top arrow follows. D 

Corollary 3.37. Kq{(3) is surjective. 

These observations are summarized in the following enhancement of fl3.90p 

Ko{R{y) ® R{y') ® Ha) = ^f. ® a^w 









Surjectivity of ^ 


3.8.4 Idempotents in U 

Let 1 = ei + ■ ■ ■ + Cfc, respectively 1 = e'^ + ■ ■ ■ + e'y, be a decomposition of 1 G 
End^(£^,ylA) = R{i^)o, respectively 1 G Endi^^S^^/lx) — R{^')o, into the sum of minimal 
mutually-orthogonal idempotents. Here -R(z/)o denotes the degree subalgebra of R^u). 
Each term in the decomposition 

k k' 
r=l r'=l 


of 1 G R{i^) ^Ri^u') Ha is a minimal degree idempotent, in view of discussion preceding 
Proposition 13.351 


Cr^r' ■■= a [tr ® c'., O 1) G End^(£:^ _^.1a) (3.101) 

be the corresponding idempotent in the endomorphism algebra of £^i^ _,^'1a which may not 
be minimal. We can decompose 




into a sum of minimal mutually-orthogonal degree zero idempotents try^r" € E'^'^u^f-v-v'^x) 
Homomorphism (5a induces a surjection of Grothendieck groups and maps each min- 
imal idempotent e^ ® e'^, ® 1 either to or to a minimal degree idempotent in Ru-u',x 
(Proposition I3.36J ). Consequently, for each {r,r',r") the image /3{eryy) is either or a 
minimal idempotent in Ry_yi\. Moreover, for each {r^r') at most one of f3{eryy) 7^ in 
Ri,-u\\- We can relabel idempotents so that /3(ej.,r',i) 7^ and P{eryy) = for r" > 1 
whenever Pa{er ® ej,, ® 1) 7^ 0. Necessarily, P{eryy) = for all r" if Pa{er ® e^, (g) 1) = 0. 
If P{er ® e^, (8> 1) = then Cryy G Iu-u',\- A homogeneous element a G Iu,-u',x can be 
written as a finite sum a = Yl^=i (^'s^s-, where a^, a'^ are homogeneous. 

a'^ G U*{Si(s)l\,Sy^^i,l) 


as G U* {Sy^^ylx, Si(s)l\) , 

and i(s) G SSeq with || i(s) || < || ^^ || + || J-'' \\- Indeed, an element of Xj^ _i//,a can be 
written as a linear combination of diagrams with fZ-turns. Cutting each diagram in the 
middle allows us to view it as composition 




with the length || i(s) \\ of the sequence i(s) strictly less than the sum of lengths 

+ II ^' II- 


Surjectivity of ^ 52 

Choose such a decomposition for each 


dr^T' ,r" £ 1u,—u',\) (ir,r' ,r" ^ / CLgCLg, (3.104) 


where in the notations we suppress dependence of u{r, r', r"), as{r, r', r") and a^(r, r', r") on 
the three parameters. Multiphcation by er^r',r" is the identity endomorphism of {Ey^^yilx, Cryy). 
We can view this indecomposable 1-morphism of U as an indecomposable projective mod- 
ule (call it P) over the graded idempotented ring ^U^, ft = X + ux — ^'x- Then the identity 
endomorphism of P factors through projective module Q corresponding to the direct sum 


for some tg G Z. Therefore, P is isomorphic to a direct summand of Q, and the 1-morphism 
{Su-u''^x,^r,r',r") oi U is isomorphic to a direct summand of ^^^iSi(s)^x{ts}- 

Define the width || P \\ of an indecomposable 1-morphism P G Ob(W(A,/i)) as the 
smallest m such that P is isomorphic to a direct summand of Silx{t} for some i G SSeq, 
II i 11= m and t G Z. 

For example, if P has width 0, then P is isomorphic to a direct summand of 1a{0 
for some A and t. One-morphism 1a{0 is indecomposable, since its endomorphism ring 
U{l\{t},lx{t}) = k, or (a possibility if the calculus is degenerate). This implies that 
any width zero 1-morphism is isomorphic to 1a{0- 

Lemma 3.38. If P has width m then P is isomorphic to a direct summand of Sy^-y'lx{t} 
for some u, z/' G N[/], || z^ || + || z^' ||= m, A G X, and t G Z. 

Proof. Ifijj G I audi = i'—i-l-ji" has length ?7i, then Pis direct summand of £^jlA{t} if an 
only if it is a direct summand of Si'^j_ii"l\{t}. Indeed, these two 1-morphisms are either 
isomorphic (if i ^ j) or differ by direct summands Si'i"l\{t'}, all whose indecomposable 
summands have width at most m — 2, and thus cannot be isomorphic to P. By assumption, 
P is isomorphic to a direct summand of Sil\{t} with || i \\= m. Moving all positive 
terms of i to the left of all negative terms produces a sequence j{—j') with j,j' positive, 
II J II + II J' \\— ""^ ^^d -P being a summand oi Sj(^^j^l\{t}. But this 1-morphism is a direct 
summand of Si^^^iyrlx{t} with u, respectively z/', being the weight of j, respectively j'. D 

Proof of Theorem \l.l[ We show that [P] is in the image of 7: ^U -^ Kq{U) by induction 
on the length of P. Let P have length m. Then P is a direct summand of S,y^-^'l\{t} for 
z/, z/' as above. By shifting the degree of P down by t, P = {Su-u''^x,€.r,r',r") for at least 
one minimal idempotent Cryy. We must have r" = 1 and l3{eryy>) 7^ 0, for otherwise P 
is isomorphic to a direct summand of ^gSi(s)lx{ts}, and since || i(s) \\= m — 2, the width 
of P is at most m — 2, a contradiction. Thus r" = 1 and [3{eryy) 7^ 0. 
We have 

[P] = [£:,,_,,! A, e,,,.,i], (3.105) 


Injectivity of 'y in the nondegenerate case 53 



/ ^ [£u-u''i-X,er,r',r"] = [£u-u''^X,er,r'], (3.106) 


from (13.1021) . For r" > 1 we have P{eryy) = 0, and {£i,^^tj'lx,eryy') is isomorphic to 
a direct summand of a finite sum of Sil\{t}, for sequences i of length m — 2. Each 
indecomposable summand of {£u-^'l\,er,r'y) has length at most m — 2. By induction 
hypothesis, [S^^^^'lx,eryy] belongs to the image of 7 for all 2 < r" < k{r,r'). Thus, 


[P] = [Si,-yil\,ery,l\ = [Si,-yil\,ery] — 2_^ [£-u-u''^\,^ryy'] G [Su-y'l\,ery] +7(ytU). 


It now suffices to show that [£y-yilx, Cry] belongs to image of 7. But the idempotent 
Cry is the image oier®er'®l in Riy) ® Riy') ®Yi\, and the Grothendieck group of the latter 
is isomorphic to A^iy) ® A^i.^')- Therefore, [£u-u',^ry] is in the image of A^iy) ® A^i.^') 
under the composition map 

^f(i/)^^f(z/') ^1,(^U)1a ^^i^o(W(A,/i)) 

x(^y V x+y-lx, 

X ^— *> x"*" : A^i.^) ^ U^, 

y'^y'^ ■■ A^W) — ' U". 

This completes the proof of surjectivity of 7. D 

3.9 Injectivity of 7 in the nondegenerate case 

Assume that our graphical calculus is nondegenerate for a given root datum and field k, 
so that -Bjj,A is a basis of H0Mz^(£^j1a,£^j1a) for all i, j and A. Then 

gdim HOM,, {Edx, £^lx) = $^ g* dim^U{£dx{t}, £jlx). (3.107) 

Since the calculus is nondegenerate, 

gdmi}10Mu{£ilx.£jlx) = T^{Eilx.Ejlx). (3.108) 

and -EjIa, over all i. A, span U, the Q(g)-algebra homomorphism 

7q(,) : U ^ Ko{U) ®^[,,,-i] Q(g) (3.109) 



intertwines the Q(g)-seniilinear forms 7r(, ) on U and gdim HOMi^(,) on Kq. The latter 
form, which we denote (, )^, is given by 

(in [Q])^ ■■= $^g*gdim(W(P{t},Q)) (3.110) 


for any two 1-morphisms P,Q E U{X,^), and extends to the entire Kq{U) ®'&[q,q-'^] Q(q') 
via (Q(g)-semihnearity and the orthogonahty condition (a;,y)^ = for x G Kq(IA{\, jj)) , 
y G Kq(U{X', fi')) unless A = A' and fi = /i'. 

By Proposition 12.51 (, ) is nondegenerate on U. Therefore, 7(Q(g) is injective, implying 
that 7 is injective. 


4 Categorification of U for si 

4.1 Forms of quantum sin 

We consider various forms of the quantized enveloping algebra of s[„ corresponding to the 
root datum of the Dynkin graph 

12 3 n-l 

o o o ■ ■ ■ o 

For this root datum, any weight A G X can be written as A = (Ai, A2, . . . , A„_i), where 

X, = {t,X)- 

The algebra Uq(s[„) is the Q(g)-algebra with 1 generated by the elements Ei, Fi and 
i^j ^ for i = 1,2, . . . ,n — 1, with the defining relations 

K,K-' = K-'K, = 1, K,K, = K^K„ (4.1) 

K,E,K-^ = q'-^E^, K,F,Kr' = q-'-^F^, (4.2) 

where i ■ i = 2, i ■ j = —1 ii j = i ±1, and i ■ j = otherwise, 

E,F, - F,E^ = 6,, ^'~\ , (4.3) 

q — q~^ 

E^E, -{q + q-')E,EjEi + EjE^ = if j = z ± 1, (4.4) 

F^F, -{q + q-')F,FjF, + F,F^ = iij=t±l, (4.5) 

E,Ej = EjE„ F,Fj = F,F, if ^ - j| > 1. (4.6) 

Recall that U(s[„) is obtained from Uq(s[„) by adjoining a collection of orthogonal 
idempotents 1a indexed by the weight lattice X of sin, 

Ixlx' = Sxx'lx, (4.7) 


The 2-category l{^{sln) 55 

such that if A = (Ai, A2, • . . , Xn-i), then 

K,lx = lxKi = q^'lx, E^lx = lx+i^E,, Fdx = lx-ixFi, (4.8) 


{(Ai + 2, A2 - 1, A3, . . . , A„_2, A„_i) if i = 1 

(Ai,A2,...,An-2,A„_i - l,A„_i + 2) ifi = n-l (4.9) 

(Ai, . . . , Ai_i - 1, Ai + 2, Ai+i - 1, ... , A„_i) otherwise. 

The Z[g, g^^]-algebra ^U(s[„) is the integral form of U(s[„) generated by products of 
divided powers E^^Hx := §^lx, F^^hx := ^Ia for A G X and i = 1, 2, . . . ,n - 1. The 
relationships are collected below: 

^'Cid integral 

^^ / . N idempotents . form 

4.2 The 2-category U^isln) 

We introduce a 2-category U^ that is defined analogously to U in the s[„-case, but the 
i?(i/)-relations have been modified to the signed i?(z/)-relations given in [17]. Namely, the 
i?(i/)-relations in U are replaced in U^ by the signed i?(z/)-relations obtained from the 
oriented graph 

o— ^O— i^O^ ■•■ — h^o , (4.10) 

with vertices enumerated by the set {1,2,..., n— 1}, using signs t^ = Tji = —1 for all edges. 
It was observed in [TTj that the resulting ring Rrii') is isomorphic to Riy). In Section USUI 
following the definition of ZY_, we extend this isomorphism to an isomorphism U — > U^ of 
2-categories. The 2-category IA_ is more convenient for constructing a representation on 
iterated flag varieties in Section [61 

In general it is a poor practice to set up an isomorphism rather than an equivalence 
of categories, not to mention 2-categories. However, having an isomorphism lA -^ IA_ is 
justified, since U and U_ have the same objects, morphisms, and generating 2-morphisms. 

Definition 4.1. W_^(s[„) is a additive k-linear 2-category with translation. The 2-category 
W^(sl„) has objects, morphisms, and generating 2-morphisms as defined in (13.11) . but some 
of the relations on 2-morphisms are modified. 

• The s[2 relations and the shift isomorphism relations are the same as before, see 
equations fl3J|) - fl3:9|) . 

All 2-morphisms are cyclic with respect to the biadjoint structure as before, see (13.3 
and dXTU]) . 

The relations (13.131) hold. 


The 2-category W^(s[„ 


The signed -R(z^) relations are 
(a) For i ^ j, the relations 



iU-j = 0, 

iii-j = -1. 

(b) For i ^ j, the relations 

X> - X/ X 

for all A. 

(c) Unless i = k and j = i ±1 

For i ■ j = i ±1 




Kl 17^ 

^ \ 

A' - K 



r ^i i i j r^ 

i / 


X' (4.12) 




The 2-category W^(s[„) 


4.2.1 The 2-isomorphism T.:V(^U^ 

Define an isomorphism of 2-categories T,: U ^ U^ on objects by mapping A 
hom categories by graded additive k-linear functors 


1-1 la 1m 



A, and on 



ll '^■a. '^rri 


■X- , 

^1 let ^a + 1 1 

if ia = ia+1, or i 


2l la ^CK + 1 "^rj 


2l ^CK ^CK + 1 "irj 





■^ otherwise 

^1 ^CK ^a + 1 ^77], 

'^ for all orientations 

2l Zq ^ck + 1 ^r) 



for all orientations 

«c« + l in 

id ^Ct + l ^n 

Above, the i^ in (—1)*° refers to the enumeration of the vertex i^ in (ll.lOp . One can 
check that the above transformations respects the s[2-i'elations and the cyclic condition. 
Furthermore, S maps the i?(i/)-relations to their signed analogs by rescaling the generators 
as above. S is a 2-functor and an isomorphism of 2-categories. 

Remark 4.2. Since the 2-categories U and U^ are isomorphic, by the universal property 
of the Karoubi envelope, their Karoubi envelopes are isomorphic as well. 

Define U^ to be the graded additive k-linear category which has the same objects and 
1-morphisms as U^ and 2-morphisms 




4.2.2 Relation to rings R{i') 

Regard the graded k-algebra -R(z^) with system of idempotents {Ij} as a pre-additive k- 
linear category whose objects are {i \ i E Seq(z/)}. The k- vector space of morphisms from 
i to i' is j'-R(z/)j. The composition j"i?(z/)j' ® iiR{v)i -^ i"R{i')i is given by multiplication 
in R{u). 

For any weight A there is a graded additive k-linear functor 


W:(A,A + z/x) 




that takes object i to Sil\, and is given on generators of horns by 





X . 

''I ^a ^CK + l '^Tl 

1-^ < 

«1 ia «Q! + 1 «n 

■X- , 

?X ?CK ^a + 1 ?-■ 

II Za — la+l, or Za 



Linear combination of diagrams in -R(z^) get sent to the corresponding rescaled hnear com- 
bination in W^(A, X + ux) with the weight A labelhng the far right region. In what follows 
we will refer to those diagrams in the image of lx as -R(i^) generators. The image of -R(i^) 
is spanned by diagrams with all strands upward pointing and no caps or cups. 

5 Iterated flag varieties 

5.1 Cohomology of n-step varieties 

The material in this section generalizes that of [211 Section 6]. The reader is encouraged 
to start there for more examples and greater detail in the 5 [2 case. We enumerate by 
/ = {1,2,...,?T, — 1} the vertex set of the Dynkin diagram of sin 

— O 

Fix A^ > 0, and consider the variety Fl{n) of n-step partial flags F 

F = (0 = Fo C Fi C . . . C F„ = C^) 

in C^. The dimensions of the subspaces Fi are conveniently expressed as a vector, 

dimF = (dim Fq, dim Fi, dim F2, . . . , dim F„). 

The connected components of Fl{n) are parameterized by non- negative integers k = 
(fco, ki, ^2, • • • , kn) such that = ko < ki < k2 < . . . < kn = N. The connected com- 
ponent Fl{k) corresponding to k consists of all flags F such that dimF = k. Throughout 
this section we refer to the terms ka of k with the convention that 


if a < 0, 

A^ if a > n. 



Cohomology ofn-step varieties 59 

The cohomology algebra of Fl{k) is Z_|_-graded, 

Hk:=H*{Fl{k),k)= H''{Fl{k),k). 


For l<j<n, 0<a<kj — kj^i, let x{k)j_a be a formal variable of degree 2a. The ring 
Hk is isomorphic to the quotient ring 

k[a;(^)j,i, x{k)j^2, • • • , x(fc)j,fe^_fe^_ J 1 /4,Ar, (5.2) 

where Ik,N is the ideal generated by the homogeneous terms in the equation 


J](l + x(i),, it + x{k),,2t^ + ■■■ + x{k),,k,-k,.,t''^-'^-') = 1. (5.3) 


Above, t is a formal variable used to keep track of the degrees. For notational convenience 
we add variables x{k)jfl and set x{k)jfl = 1. Furthermore, we set 

x{k)j,a = a a > kj — kj^i. (5.4) 

It is helpful to express the above relation in an alternative form. Let x{k)i^a denote the 
homogeneous term of degree 2a in the product 


J](l + x(fc),, it + x{k)j,2t^ + ■■■ + x(fc),- fc^._fc^_,t^^-^-^). (5.5) 

For example, if n = 4 and k = (1, 3, 4, 7), then equation (15.31) becomes 

(1 + x(fc)i,it) (1 + x(^)2,it + x{k)2,2t^) (1 + a;(fc)3,it) (l + x{k)4,it + x{k)4,2t^ + x(k)4,3t^) = 1 

and the terms x{k)2,a are given by omitting the second term and multiplying out the rest 

(1 + X(fc)i,it) (1 + X(^)3,it) (1 + X(fc)4,lt + X(fc)4,2t^ + X(fc)4,3t^) 

SO that 

xik)2,o = 1, 

x{k)2,i = x(fc)i,i + a;(fc)3,i + x(fc)4,i, 

x{k)2,2 = x{k)i^ix{k)3^i + x(fc)i,ia;(fc)4,i + x(^)3,ia;(i)4,i + a;(^)4,2, 

3^(^)2,3 = 3^(^)1,1^(^)3,12^(^)4,1 + 3^(^)1,12^(^)4,2 + 3;(^)3,ix(^)4,2 + a;(fc)4,3, 

a;(^)2,4 = a;(fc)i,ix(fc)3,ix(fc)4,2 + a;(fc)i, 1^(^)4,3 + a;(fc)3,ia;(^)4,3, 

x{k)2,5 = a;(fc)i,ix(fc)3,ix(fc)4,3, (5.6) 


Cohomology ofn-step varieties 60 

and x(fc)2,a = for a > 5. It is clear that (15. 3p can be written 

Yl ^(^)jJ ^ik)j,a~f = 5a,o (5.7) 

for any I < j < n, where 5afl is the Kronecker delta. We call the elements x{k)j^a dual 
generators in light of (15.71) . 
For 1 < i < n — 1 let 

+ik = {{+ik)o, {+ik)i, {+ik)2, ..., i+ik)n), = {+ik)o < {+ik)i < ■ ■■ < {+ik)n = N, 

be the sequence obtained from the sequence k by increasing the ith term by one 

+ik := (fco, ki, ^2, • • • , fci-i, ki + 1, ki+i, ...,kn) (5.8) 

if fcj + 1 < fcj+i, or by setting the sequence to the empty sequence if fcj = fcj+i. Namely, 
{+ik)j = kj a j j^i and {+ik)i = ki + 1 ii ki + 1 < fcj+i. 

When ki + 1 < ki^i then H^.^ is the cohomology ring of the partial flag variety consisting 
of flags F with dimF = +jfc. The ring H^^k is given by 

H+rk = I 09 H^i+iI^)jA^ x{+ik)j,2, . . . , x{+ik)j^(^^^k),-{+,k),.A ) /I+,k,N 

= (^ k[x{+ik)j^i, ..., x{+ik)j^kj-k,^i] ® k[x{+ik_)i^i, ..., x{+ik)k,-k,_i+i] 

®k[x(+i^)i+i,i, • • • , 3;(+ifc)fc,+i-fe,-i]//+,fc,7v (5.9) 

where Ij^^k^N is the ideal generated by the homogeneous terms in 

\{[Y,<+^k),,aA =1. (5.10) 

i=l \a>0 / 

We define i/© = 0. 

Going back to the example of ri = 4 and k = (1, 3, 4, 7), then +3^ = (1,3, 5, 7), so that 

H^^k = lk[a;(+3^) 1,1, ^(+3^)2,1 ) a;(+3fc)2,2, 2;(+3i)3,i) a;(+3^)3,2, a:(+3^)4,i, a:(+3^)4,2]//+3fc,7 

where Z+gfe.T is the ideal generated by the homogeneous terms in 

(1 + a;(+3^)i,i)(l + x{+3k)2,i + x(+3^)2,2)(l + a;(+3fc)3,i + ^(+3^)3,2) 

X (1 + a;(+3fc)4,i + a;(+3^)4,2) = 1- (5.12) 

Similarly, we write _jfc = (/cq, ki, k2, ■ ■ ■ , h-i, ki — 1, ki^i, . . . , kn) for the sequence k 
where we have subtracted one from the ith position whenever ki^i < ki—1. The cohomology 
ring of the fiag variety Fl{_ik) is H-^., which can be expressed explicitly in terms of 
generators as above. When A;j_i = ki then _jA; = and H_-k '■= 11% = 0. 


Cohomology ofn-step varieties 61 

5.1.1 Flag varieties for the action of Ei and Fj 

For 1 < i < n — 1, define 


+i _ ) ("^0) kl, /C2, . . . ki, ki -\- 1, Kj+l, . . . , kn-l, kn) II ^j+l > ^j + 1, 


(fcO) "^1; "^2? • • • ki-l, ki i, ki, . . . , kn-l, kn) II fcj i ^ ^i-l; 


For ^ * 7^ the variety -F/(A^ *) is the component of Fl{n + 1) consisting of flags F 
such that dim F = k ^. The cohomology ring of Fl{k *) will be denoted by Hi^±i. The 
cohomology ring Hf^+x is 

i7fc+, = (g) k[x(^+%- 1, . . . , x(^+%- fc^,_fe^._J®k[ei]®k[a;(^+')i+i,i, . . . , a;(^+*)i+i,fc^^^_fc^„i]/V._^ 

where /;!.+* ^ is the ideal generated by the homogeneous terms in 

(1 + e.t) (1 + a;(A;+*),+i,it + ■ ■ ■ + x(i+^),+i,fc,^,_fc,_it^»+^-^-^) (5.14) 

X n ( E ^(^^^)^-./ ^0 = 1- 

jVi+i /=o 

The forgetful maps 

F/(A;) ^^^ F/(i+^) ^^ F/(+,^ ) 

induce maps of cohomology rings 

Hk — ^ Hj^+i ^-^— H^^k 

that make -f/^fc+» a right Hk Cg) if^.fc-module. Since the algebra H^.^ is commutative, we 
can turn a right iJ^^^-module into a left iif_^^fc-module. Hence, we can make Hf^+t into a 
(if^.fc, iJfe)-bimodule. In fact, i^fc+» is free as a graded iffc-module and as a graded H^.k- 

These inclusions making -f/^^+i a [H^^k, -f/fc)-bimodule are given explicitly as follows: 

Hk ^ > H^+^ (5.15) 

X{]^j,a ^-^ x{M^^)j,a for j 7^ i + 1 , 

x{]^i+i,a ^ ii ■ a:(^^*)i+i,a-i + a;(fc+*)j+i,„ , 


H^.k ^ ^ H,^. (5.16) 

a^(+i:^)i,a ^ a;(^+*)j,a for j ^i , 


Graphical calculus for iterated flag varieties 62 

Notice that x{k)j^a and x{+ik)j^a for j j^ i,i + 1 are mapped to the same element of Hk+i- 
Using these inclusions we identify these elements of Hk and H^.k with their images in the 
bimodule Hf^+i. Furthermore, we can also express the generators x{k'^^)i^a and x{k^^)i+i^i3 
of Hf,+i as the images of certain generators in Hk or H^^k- Thus we can write Hf^+i as 

Hk+^ = (^ k [x{k)j^i, ..., x{k)jMj-k,_^] ® k[x{+ik)i+i^i, . . . , x{+ik)i+i^k,+-,-k,-i] ® k[^i]/4+.,Ar 

or equivalently, 

Hk+^ = (^k [xi+ik)j^i, . . . , x{+ik)j^kJ-k,^^] 0k[x{k)i^i, . . . , x{k)i^ki-k,^^] ® k[6j]/4+',iV' 

where /^+. jy is the ideal described above. Therefore, ^j is the only generator of Hf.+i that 
is not identified with a generator of Hk or -ff+^fc under the above inclusions. 

Remark 5.1. The generators x{k)j^a, x{+ik)j^a, ^i of Hj^+z correspond to Chern classes 
of tautological bundles over the variety Fl{k^^). The generator ^j corresponds to the line 
bundle Fk+i/Fk- associated to the subspaces Fk C -F^ +1 created by the subspace insertion 

Definition 5.2. The set of multiplicative generators 

x{k)i^ai for < ai < ki- h^i, 

x{+ik)i+i^ai+i for < ai+i < ki+i - ki - I, 

x{k)j,a = x{+ik)j^a £ Hi^+i for j ^ i,i + 1, and < a^ < kj — fcj-i, 

for the ring H^+i, corresponding to the Chern class of the tautological line bundle, and 
to the canonical inclusions flS.lSp and fl5.16p of generators in H^.k and Hk into if^+i, are 
called the canonical generators of Hf,+r. 

Using commutativity we can regard if^+« as an {Hk, if ^.fc) -bimodule. The generators 
of Hk and H^.^ that are not mapped to canonical generators can be expressed in terms of 
canonical generators as follows: 

x[k_)i^i^a = Si ■ x[^ik_)i^iiy^i + x(-|-jA;jj-|-i^Q, 

for all values of a. 

5.2 Graphical calculus for iterated flag varieties 

5.2.1 Rings Hk 

All of what has been described can be easily visualized using a shorthand notation in which 
the generators of Hk are drawn as labelled bubbles floating in a region carrying a label A 


Graphical calculus for iterated flag varieties 63 

called the weight. The label A will be important when we relate partial flag varieties to 
the 2-category U categorifying ^U(s[„). More precisely, A will correspond to a weight of 
the irreducible representation of yiU(s[„) with highest weight (A^, 0, 0, . . . , 0). 

To a sequence k = (fco, ^i, ^2, • • • , kn) as above associate A = X{k) = (Ai, A2, . . . , A„_i) G 
2"^i where 

Aq, = -ka+i + 2ka - ka-i. (5.18) 

The weight corresponding to the sequence +jfc is defined analogously, where ki is replaced 
by ki + 1. Comparing with (14.91) it is clear that \{+ik) = X + ix- Similarly, X{~ik) = X — ix- 
With this convention, the generators of H^ and H^^k are depicted as follows: 


x{k)j,^:= {JB (5.19) 

X + ix 
x{+ik)j,a:= [B (5.20) 

where the identity is depicted by the empty region of the appropriate weight. Products of 
generators are depicted by a bubble in the plane for each generator present in the product. 
Diagrams are only considered up to planar isotopy. A generic element in Hk can be depicted 
as a formal linear combination of such diagrams. For example, if n = 4, fc = (1,3,4,7) 
then the element x{k)iix{k)i^^ + 5 • x{k)2,^i e H^ is represented as 

A A 

® +^ [£3 

If we depict the dual generators x{k)j^a of Hk defined in ( 15.50 as 

a;(fc)j,a = 03 

then the defining relations (15. 7p for H^ become the equations 

a X a X A 

Y, E GE3 = E [^^ GB = ^^.o (5-21; 

/=0 3=0 

for a > and 1 < j <n. Notice that 

A A A 

(To] = (To) =1, [73 =0 if a < 0, or A;^ - %_i < a, 

see ( 15.41) and the comments preceding (15.41) . 


Graphical calculus for iterated flag varieties 


5.2.2 Bimodules H^.+^ 

The identity element in ifj^+» is represented by a vertical line labelled i 

Hu+i 3 1 

\ + ix 


where the orientation indicates that we are regarding if^+, as an {Hj^ik^ -f4)-bimodule. The 
A on the right hand side is the weight corresponding to k. Hence, having A on the right 
hand side of the diagram indicates the right action of Hk on H^+t. Similarly, the X + ix 
on the left indicates the left action of H^.^ on H^+i . 

When regarding if^+« as an {Hk, -ff+ifc)-bimodule we depict it in the graphical calculus 
with the opposite orientation (a downward pointing arrow). 

H.+, 3 1 


X + ix 

We will often omit the weights from all regions but one, with it understood that crossing 
an upward pointing arrow with label i from right to left changes the weight by ix, and 
crossing a downward pointing arrow from right to left changes the weight by —ix- 

Equations (15.171) show that all of the generators from Hj^+z except for C,i can be in- 
terpreted as either generators of Hk or H^^^ under the natural inclusions. This fact is 
represented in the graphical calculus as follows: 

03 jV ^ + 1 (5.22) 

f/fc+> 3 x{+ik)j^i3 :-- 


3 ^i 



where each diagram inherits a grading from the Chern class it represents (degx(^)j_Q, = 2a, 
deg a;(+jfc)j,/3 = 2/3, and deg^i = 2). Equation (I5.22p is meant to depicts the generator 
x{k)j,a G Hi.+i as the element x{k)j^a ^ H^ acting on the identity of -f^^+i. Likewise, the 
generator x{+ik)j^^ G Hf^+t is depicted as the element x{+ik)j^j3 G H^.^ acting on the identity 
of Hp.+z. The generator ^j is represented by a dotted line so that ^f is represented by a 
dots on a line, but for simplicity we write this using a single dot and a label to indicate 
the power. 

The identification (I5.15P and (I5.16P of x{k)j^, 
leads to the graphical identity: 

with x{^ik)ja in H^+t for j ^,i,i + 1 





for j ^ i,i + 1. 



Graphical calculus for iterated flag varieties 


Similarly. (5.17P provides the identities 

i A 




i+l,a-l] !' + i+l,Q! ,. 
i % 

i \ A 

, , i.Q— 1 -|- , V i,a 


expressing non-canonical generators in terms of canonical generators. It is sometimes 
helpful to express the canonical generators in terms of non-canonical generators: 



.f=o I 


^ A 





/ A 


which can be verified using (15.261) and (15.271) . From these equations we can derive other 
useful identities: 



I a A 


Proposition 5.3. 













/ A 

if j = i 
if j = i + 1 



Graphical calculus for iterated flag varieties 





E?.„(-i)' I' 




iij = i 

if j = i + 1 


Proof. Recall that the elements x{k)j^a are sums of homogeneous symmetric terms in all 
variables except for x{k)j^fi. By f l5.25p all terms x(fc)^,a for £ 7^ z, z + 1 can be slid across the 
line labelled i. The case when j 7^ i,i + l then reduces to the problem of sliding symmetric 
combinations '^f=Qx{k)ijx{k)i+i^/3^f across a line labelled i. Such slides are determined 
by the following calculation in Hj^+i: 

/3 /3 / 

/=o /=0 9=0 

/3 / 


/=0 3=0 

13 f 

/=0 3=0 


Change variables to /' = / + 1, and g' = g + 1 in the first summation, so that all terms 
cancel except for g = Q term in the second summation; this term is equal to 



Hence, when j 7^ i,i + l the elements x{k)j^a and a;(+jA;)j „ can be slid across a line labelled 

The dual generator x{k)j^a contains symmetric homogeneous combinations of variables 
x{k)i^i3 for i 7^ j. When j = i in (I5.32p all terms in x{k)i^a slide across lines labelled i 
except for the variables x{k)i+i^i3. Using (I5.26P to slide these across establishes (I5.32p for 
the case j = i. Similarly, when j = i + 1 all terms in x(A;) j+i_q, slide from right to left across 
lines labelled i except for the variables x{k)i^i3. Using (15.291) to slide these completes the 
proof of (15.321) . Equation (I5.33P is proven similarly. D 


Graphical calculus for iterated flag varieties 


By duality, analogous equations as those above hold for downward pointing arrows. For 
example, equation fl5.30p implies 





5.2.3 Bimodules Hj^i 

For the remainder of this paper we write a signed sequence i = £l^l£2^2 • • • ^m'im as i = 
S1S2 ■ ■ ■ Sm with Sa = Saia- For Sq = Soia Write sak. for the sequence obtained from 
k = {ko, ki, ^2, • • • , kn) by increasing the ki^ by 1 if ^q, = + and ia < ia+i, decreasing the 
sequence by 1 if £„ = — and i^-i < ia, and setting the sequence to otherwise. Then 
ik is either set to 0, or else it denotes the sequence obtained from k with the ki^th term 

increased by one if Sa = +ia, or decreased by one if Sc 

sequentially for each Sa in 

the signed sequence i, reading from the right. It is clear that the sequence ik is equal to 
the sequence jk whenever i,JE SSeq with ix = j x ^ ^ and ik^% and jk 7^ 0. 
We write Hf^i for the {H.k, -f/fc)-bimodule 

Hui '■= H, 


1 "• 

St-n "- — 


This bimodule can also be described as the cohomology ring of the variety oi m + n 
step iterated partial flags corresponding to the sequence obtained from k by the ordered 
insertion of subspaces determined from the signed sequence i. 

The {H.k, iffc) -bimodule Hf,i can be understood using the graphical calculus. A general 
signed sequence i = SiS2---Sm, is represented by a sequence of lines coloured by the 
sequence iii2---im, where the line coloured by ia is oriented upward if s^ = +ia and 
oriented downward if s„ = —in- 


1. For the signed sequence +j + i E SSeq consider the {H^^^^k, -f/fc)-bimodule 




,k+i ^^+,fc ^fc- 


As explained in the previous section, the identity elements of the {H^.k, -f/^) -bimodule 
Hf^+i and the {H^.^_.k, if^.fc)-bimodule H .).] are depicted as 

\ + ix 



X + ix+ jx 

\ + ix 



The identity element of Hk+^+* i^ represented by the diagram 

X + ix+ jx 

\ + ix 




Graphical calculus for iterated flag varieties 


The region in the middle of the two hnes is labelled by the weight X+ix corresponding 
to +jfc. The tensor product over the action of H^.k is represented diagrammatically 
by the fact that a labelled bubble in the region with weight X + ix can be equivalently 
regarded as an element of H^.k acting on the line corresponding to Hf^+i, or the line 
corresponding to H f,j. 

Action of Hk 

Action of H^h Action of H^^h 





J i J i J i 

The weight A on the far right, and the weight oi X + ix + jx on the far left, indicate 
that this diagram is describing an {H_^_j^^k, -f/fc)-bimodule where the various actions 
are depicted as above. 

2. For i = +ii + 22 + ^3 ■ ■ ■ + V the identity element of the bimodule H/^i is depicted by 
a sequence of upward oriented labelled lines 

Hk' 3 1 

/ \ / V / V 

^l 12 13 

/ V y V 



^m — 1 In 

The canonical generators ^i^ of each term in the tensor product (15.371) are represented 
graphically by a dot on the line labelled ia- 

H^i 3 1^1 


/ \ / \ 




«1 «2 la Im 

k is represented by the regions 

Likewise, the tensor product over the rings if, 

between lines. Again, the weights on the far left and right of the diagram indicate 

the bimodule structure. 

Action of Hk 

Action of H 



/ V / V 



/ \ / \ 

Action of H.k 




l\ 12 


11 12 

3. Consider the {Hk-, iife)-bimodule corresponding to the tensor product 

Hk+' ®// . H^+i 


where in the first factor we are regarding H^+t as a {Hk-, -ff^,fc)-bimodule. The identity 
element of this bimodule is represented by the diagram 


X + ix, 




Graphical calculus for iterated flag varieties 


5.2.4 Identities arising from tensor products 

Bubbles with a given label (j, a) can pass from right to left, or left to right, through a line 
coloured by i as long as j y^ i,i + 1. Iij = ioTJ = i + l then a bubble can move through 
a line subject to the rules fl5.26p -( l5.29p . Furthermore, dots on a line can be exchanged for 
bubbles in the neighboring regions using (15.301) and (15.311) . Dual bubbles corresponding to 
dual generators can be slid across lines using the rules (I5.32p and (I5.33p . 

The following Lemma is needed for the definition of the 2-representation F given in the 
next section. In particular, parts (i) and (ii) are used to provide two equivalent definitions 

and (iii) and (iv) are used to show that 

^(Va) ^ ^(Va)' ^(Va) ^ KVa)- 

Lemma 5.4. The following identities hold: 

i) Equivalent definitions of cups labelled i: in the ring if^+» ®h ^ H^+z we have 

^ 1 


E(-ir^ ' 

. i,a-f 



for all a e N. 
ii) In the ring H_j.+i ®h .^ H_.j^+t we have 


E(-i)°-^ Q^ I I 

9 A 






^(_l)a-9 [i+l,a-g] 

I t 


E(-l)"-^e/®a;(A;),+i,„_/ = E(-ir~^a;(^),+i,„_,®ef 

/=0 3=0 

for all a e N. 
iii) Dot slide formulas for cups: in the ring HJ.+^ ®h .^ H^+i we have 


''-f~\^lf.~k.^.-9 ^9 ^ a:{k),^,,,^.,^_,., ■ 6. 




Graphical calculus for iterated flag varieties 


iv) Dot slide formulas for cups: In the ring H_.^+i ®h .^ H_.^+^ we have 



Proof. Part i) follows from the chain of equalities: 


/=0 9=0 


If we re-index by letting f = a — f and switch the order of summation we have 

a a—g 

9=0 /'=0 




i lOTI i 


o "^a 


which after re-indexing completes the proof of Part i). Part ii) is proven similarly. Part iii) 
follows from Part i) by letting /' = / + ! and adding and subtracting the term (— l)""''^! §5 
x{k)i,a+i, SO that 

Y.^-ir'Ur ® <k) 




= J](-l)-+i-^' i( ®x{k\^+^_f, - (-l)"+il ®x(A;),,„+i 

Part i) Y^{-lY+^-H{k\^+^^g ® ef - (-l)"+ll ® X(fc),,,+i. 

Pulling off the g = Q term and re-indexing g' = g — 1 we have 



But the term with the summation is equal to 


9 =0 


by Part i), and the remaining terms {—'\-)°''^^ {x{k)i^a+i ® 1 — 1 ® x{k)i^a+i)i are zero when 
a = ki — ki_i since x{k)i^a+i is zero for a > ki — ki+i, see (15.41) . Part iv) is proven similarly 
using Part ii). D 


The 2-category Flag^ 71 

Corollary 5.5. The assignments (see Definition 16. ip 

r (^ J : Hk ^ (i/,+. (Sh^^, H^+.^ {1 + A;,_i - A;,+i} 


^ (^.A 

: Hk — > [H^,k+^ ®H_^k H_^k+^) {1 + ki~i - h+i} 

1 ^ ^1: Vl)''+^-'^-^ef®a;(A;),+i,,,^,_,, 


define morphisms of graded bimodules of degree 1 + Aj = 1 — A;j_i + 2ki — /cj+i and 1 — Aj = 
1 + ki_i — 2ki + ki+i, respectively. 

Proof. For the first claim it suffices to check that the left action of each generator of Hk on 

r (\j;^ j (1) G [Hj^+i ^H k ^k+^ ) is equal to right action of this generator. The Corollary 

follows since bubbles can slide across lines by fl5.26l) - fl5.29l) at the cost of introducing 
powers of ^j on one of the tensor factors; by Lemma [5^ (iii) and (iv), factors of ^j can be 
slid across tensor factors in the above sums. The second claim is proven similarly and the 
degrees of these bimodule maps are easily computed. D 

5.3 The 2-category Flag^ 

Bim is a 2-category whose objects are graded rings, whose 1-morphisms are graded bi- 
modules, and 2-morphisms are degree-preserving bimodule homomorphisms. Idempotent 
bimodule homomorphisms split in Bim, so that any 2-representation \1/ : W ^ Bim extends 
uniquely (up to isomorphisms) to a 2-representation U — > Bim, see Section 13. 4[ 
Graded 2-homs between graded bimodules Mi and M2 are given by 

HOMBi^(Mi, M2) := HomBi^(Mi{t}, M^). (5.49) 


Let Bim* be the 2-category with the same objects and 1-morphisms as Bim and 2- 
morphisms given by 

Bim*(Mi, M2) := }iOM^^^{M,, M^). (5.50) 

We now define a sub 2-category Flagjy of the 2-category Bim for each integer A^ G Z_|_. 

Definition 5.6. The additive k-linear 2-category Flag^ is the idempotent completion 
inside of Bim of the 2-category consisting of 

• objects: the graded rings Hk for all k = {ko, ki, k2, ■ ■ ■ , A;„) with < A;i < A;2 ■ ■ ■ < 
kr,. = N. 



• morphisms: generated by the graded (iffc,i7fc)-bimodule Hk, the graded {H^-k,Hk)- 
bimodules if^+i and the graded {Hk,H ^.k)-him.odu\e Hj^+i for all i G /, together with 
their shifts Hk{t}, Hi^+i{t}, and Hf^+i{t} for t G Z. The bimodules H^ = -^^{0} are 
the identity 1-morphisms. Thus, a morphism from H^ to H.k is a finite direct sum 
of graded {H.^, iffc)-bimodules of the form 

for signed sequences j = siS2 ■ ■ ■ Sm with ix = j x ^ ^■ 

• 2-morphisms: degree-preserving bimodule maps. 

There is a graded additive subcategory Flag^ of Bim* with the same objects and 
1-morphisms as Flag^, but with 

Flag^(Mi, M2) := Flag^(Mi{t}, M2). (5.51) 


In Section [6] we show that Flag]^ provides a 2- representation of W^; using the isomorphism 
S : W* — >■ U^ and restricting to degree zero 2-morphisms, the subcategory Flagjy provides 
a 2-representation of U. 

6 Representing U*^ on the flag 2-category 

In this section we define for each positive integer A^ a 2-representation r^v: W^ -^ Flag^. 
The 2-functor Tx is degree preserving so that it restricts to a weak 2-functor T: U ^ 
Flag^. We will sometimes shorten Fat to F for simplicity. 

6.1 Defining the 2-functor T^ 

On objects the 2-representation Tx'- U'^ -^ Flag^ sends A = (Ai, A2, . . . , A„_i) to the rin^ 
Hk when A = A(fc), i.e., when A^ = —ka+i + 2ka — ka-i- 

Tx:U_ -^ Flag^ 

Hk if A = A(^) 

Morphisms of U_ get mapped by F^r to graded bimodules: 

Tx-.U^ -^ Flag^ 

Hk{t} if A„ = -ka+i + 2ka - ka-l 



J , . Hf.+i\t + 1 + k.i_i + ki — /Cj+ij it Aq — —ka+i + 2ka — ft^a-l 

i+,lAit| ^ - Otherwise. 


Hk-i{t + I - ki} if Xa = -ka+i + 2ka - ka-1 


DeBning the 2-functor T 



Here Hf^-i{t + 1 — fcj} is the bimodule Hf^-i with the grading shifted by t + 1 — fcj so 

{Hk+^{t + 1 - ki})j = {H^+r)j+t+i-k,- 

More generally, the l-morphism 

is mapped by Fat to the graded bimodule H^^^ with grading shift {t + t'}, where t' is the 
sum of the grading shifts for each terms of the composition in (16.11) . Formal direct sums of 
morphisms of the above form are mapped to direct sums of the corresponding bimodules. 

6.1.1 Biadjointness 

Definition 6.1. The 2-morphisms generating biadjointness in U^ are mapped by F to the 
following bimodule maps. 






n /\^ 

Hk — > f -fffc+» ®^+,fc H^+^) {1 + ki^i - ki+i} 

* j ~ 1 




''"^-\_l)fc,:-/c.-l-/ ^/ ^ x{k\k.-k.^,^f 

(^_,fc+> ^H_^, H_^k+) {1 + ki-i - h+i] 






^ 1-^ (' — 1 \ai+Oi2+'i-+ki-ki+i y / . , , , , , , , , ^ 

'^ I -LJ T X |t+l,ai+a2+l+fci-fc^+i j /j^ .x 




+1 X x(A;)j+i^Qj+Q2+l+fc»-fc«+i 

(^_.fc+« ®/^_,fc ^_,fc+») {1 + ^*-i - ki+i} -^ Hk 



2 A 





X I j,ai+a2+l+fci-i— fc, 

J (6.5) 

e ® C ^ (-l)-i+"2+i+'=»-i-'=^ X x(A;),«,+«,+i+,^_,_,^ 


DeBning the 2-functor T 



Corollary 15.51 shows that the cups above are bimodule maps. It is clear that the caps 
are bimodule maps. These definitions preserve the degree of the 2-morphisms of W^ defined 
in Section 13.11 By Lemma 15.41 the clockwise oriented cap and cup have degree 1 — Aj and 
the counter-clockwise oriented cap and cup have degree 1 + Aj so that these assignments 
are degree preserving. 

6.1.2 R{i^) generators 

Definition 6.2. The 2-morphisms f and * in U* are mapped by Tn to the graded 
bimodule maps: 



X + ix' 

- A 


i J 

Hf.+^{l + ki_i-ki+i} -^ Hl^+^{l + ki_l-ki^l} 

-Q + l 


( J 

A ~ 

'X + tx 


i J 

H^+i{l + ki_i - kij,i} -^ H^+^{l + ki^l-kiJ^l} 





Note that these assignment are degree preserving since these bimodule maps are degree 2. 
The 2-morphisms ^^ . . ^ and \C ^ ■ ^ are mapped by T to the graded bimodule maps: 


H+jk+' ®H^^k Hk+^ -^ ^+rk+^ ®^+.i ^k+' 


iT ® C^ 

e®er ^ <^ 


ai — 1 





"2-1 , -, 

if 2 ■ j = 
if z = j 




if <^^^ 


^^_,fc ^k-^ 


,k-^ '^H_ 




-_Ci2 fr^ f Ol 

«2-l , 1 r r ai-1 , , 

/=0 9=0 

if i ■ J = 
^i if ^ = J 





It is straightforward to see that these assignments define bimodule maps of degree —i ■ j. 
In the case when i = j the bimodule map is just the divided difference operator acting on 
6 and ^j. 


Checking the relations ofU 


6.2 Checking the relations of U^ 

In this section we show that the relations on the 2-niorphisnis of W^ are satisfied in Flag^, 
thus estabhshing that F is a graded additive k-hnear 2-functor. From the definitions in 
the previous section it is clear that F preserves the degrees associated to generators. For 
this reason, we often simplify our notation in this section by omitting the grading shifts 
{t} when no confusion is likely to arise. 

Proposition 6.3. F preserves the sli-relations of W^. 

Proof. The proof in [21] that the sli-relations of U^ are preserved by F generalizes imme- 
diately with only minor refinements to grading shifts and summation indices. The proof 
only makes use of fl5.2ip and (I5.25p - fl5.33p . For example, to prove biadjointness we must 

/ X + ix 




\A + 2x 



\ + ix 







\ + ix 












for all strings labelled by i. We will prove the first equality by computing the bimodules 
maps on elements ^f G Hj^+i. 

/ X + ix 



1=0 y ' 

l,Ki fei — 1 / 


Yl (-1)° \i,a-{ki-ki.i-f)] 


l.k,i ft-'i— 1 / 

A ™ „ 


since the sum can be taken to a by removing terms that are equal to zero when a < 
ki — ki_i, or by adding terms that are equal to zero if a > /cj — ki^i- The others are proven 
similarly. D 


Checking the relations ofU 


Lemma 6.4. 





\/ ^^X + tx 




. J 


X + ix 










^.-..■..fc®H,..,„fc ^^,fc + « 


-fft,+» (K'H',, H, 


fffc -^fc+i 

er ® er^ 

02 <:>, f'^i 



These biinodule maps have degree zero for all i, j G / and all weights A. 

Proof. We compute the bimodule maps directly using the definitions in the previous sec- 
tion. The case when i = j appears in [21j so we will omit this case here. The map in 
(16. lip , using (I5.43P for the cup, is given by 






X + ix 





01 ,^ £-02 


\^ (_l)"2 [i,fc,~fc,,i-/] t I [»,a2-(fc,-fc,^i-/) j 

i 3 


if z ■ J = or o — > — 6, and 

i^i "'i— 1 -I- 

\^ (-l)"2 [Xfc7-fc.-l-l-/ 


<^i "'i — 1 ~"-l- 




A + zx 

« J 



i,a2 — (ki—ki-f-i 


A + ix 

« J 



if o — > — 6. Careful calculation, keeping in mind the weights of each region, will show that 
these maps have degree zero. In both cases, the dual generators can slide across the line 
labelled j via (15.331) . After changing indices, equations (16.131) and (I6.14p both become 




,«! X + i' 



Checking the relations ofU 


where (+jfc)j — (+j/i;)j_i are the (i — l)-st and i-th terms in +jk. By adding or removing terms 
that are equal to zero, depending on whether 02 is greater than or less than (+jA;)j — (+jA;)j_i, 
the above summation can be taken to 02, so that (I5.3ip completes the proof. D 

Proposition 6.5. The equality 

r A 

i^ Vi 


r /— 





' H' 



^^ i 



of graded bimodule maps holds in Flag^ for all weights A. 

Proof. By direct calculation, using Lemma 16.41 we have that both the far left and right 
bimodule maps are given by 

^r ® ef 

"2-1 , 1 r J. ai-1 

/=0 g=0 

(ef ®e){-i} 

if i ■ J = 
if i = j 



if 6^^ 

which agrees with T { \( . . , , . , . , . 
Proposition 6.6. The equalities 








of graded bimodule maps hold in Flag^ for all weights A. 
Proof. This follows immediately from Lemma [6.41 
Proposition 6.7. The equalities 

H X/) -( X^)^ -( x^ 

hold in Flag^ for all weights A. 

Proof. Using the definition of the bimodule map r( ^>^ . ) in (16. 8p the proposition is 
easily verified. D 

r( Xa) (6«) 


Equivariant representation 


Proposition 6.8. The equalities 




iii-j = 

iii-j = -l 



X \ 


^ A 



ir\k / 








-1 (6.21) 

of bimodule maps hold in Flag^ for all weights A. 

Proof. To prove the proposition we compute the bimodule maps in (16.191) on the elements 
of the form ^"^ ® ^°'^. Bimodule maps in (16.201) and (I6.2ip are computed on elements 
^f^ ® ^"^ (S) ^^^ and ^f^ (S) ^"^ (S) ^f^, respectively. The action on other elements in the 
cohomology rings is determined by the fact that the maps are bimodules morphisms. 

The proof of these remaining relations is the same as the proof that rings -R(i^) act on 
Voiu (notation as in pjj). Replacing the variables Xk{i) G Voit with the Chern classes of 
line bundles C,k in the corresponding cohomology rings turns formulas in [161 Section 2.3] 
for the action of dots and crossings into formulas (16. 6p and (16.81) . with the signs taken into 
account. D 

Thus, we have proven the following Theorem: 

Theorem 6.9. Tn: W^ —>■ Flag^ is a 2-functor and a 2-representation. 

6.3 Equivariant representation 

6.3.1 Reminders on equivariant cohomology 

The GL(iV)-equivariant cohomology of a point [H] is given by 

Hh L{N) (pt) = H*{Gr{N, oo)) =k[xi,X2, . . . ,XN,yi,y2, ■ ■ ■]/In,oo (6.22) 

where In,oo is the ideal generated by the homogeneous terms in 

(1 + xit + X2t^ + ■■■ XNt^) (1 + yit + y2t^ + ■■■ %t-'' + ■■■) = 1. 


Equivariant representation 79 

Thus, H^^j^Jpt) is isomorphic to the polynomial ring 

H*GL(N){P't) = k[xi,X2,...,XN-l,XN] (6.23) 

with Xi in degree 2i. 

Given a sequence k = (/cq, ki, k2, ■ ■ ■ , kn) with < /ci < /c2 < ■ ■ ■ < /c„ = A^, GL(N) 
acts transitively on Fl{k), so the equivariant cohomology of Fl{k) is 

H*GLiN){Fm) = i7sU(p*) {Pt e F/(A;)) (6.24) 

where the stabilizer of a point (O C C'^^ C ■ ■ ■ C C'^" = C^) in FL{k) is the group of 
invertible block (/ci x (k2 — ki) x ■ ■ ■ x (A^ — /Cn-i)) upper-triangular matrices. This group 
is contractible onto its subgroup GL(ki) x GL{k2 — /ci) x • ■ ■ x GL{N — kn-i). Hence, 

Hk — -"GL(fci)xGL(fc2-A:i)x---xGL(Af-fc„_i)(P^) 

[a;(^)j,i, a;(^)j,2, • • • , x{k)j^k,~k,^^] (6.25) 

with dega;(^)j^a = 2a;. Thus, the equivariant cohomology of Fl{k) has the same generators 
as the ordinary cohomology ring, but we do not mod out by the ideal Ik,N- 

The equivariant cohomology rings H'^ f, and H^ j. oi Fl{^ik) and Fl{_ik) can be similarly 
computed. They have the same generators as the ordinary cohomology rings H^^^ and H.^, 
but no relations. 

The equivariant cohomology H^+i of Flik^"^) also has the same generators as the or- 
dinary cohomology ring, with no relations. Using the forgetful maps Fl{k) <— Fl{k^^) -^ 
Fl{+ik) we get inclusions 

H^ c , H^^. (6.26) 

x{k)j^a ^ x{k'^'')j^a ioT j j^i + 1 , 
a^(i)i+i,a ^ ^i ■ x{k~^')i+i,a-i + a;(fc+*)j+i,Q , 


Kk "-^ ^?- (6-27) 

x{+ik)j^a ^ a;(^+*)j> for j j^ i , 

making -f^^i a graded (i/^^^, i/jfc)-bimodule, just as in the non-equivariant case. Using 
these inclusions we introduce canonical generators of -f^f+i given by identifying certain 


Equivariant representation 80 

generators of H^ and H'-^.j. with their images in -fff+i- Thus, we can identify x{k)j^a and 


k)j^a in H^+^ when j j^ i,i + 1, and 


or equivalently, 

The generators of H^ and -f^'^^^ that are not mapped to canonical generators in H^+, 
can be expressed in terms of canonical generators as follows: 

X[k_)i+l,a = C,i • X[j^ik)ij^i^a-l + x[^ikjij^i^a 

x{+ik)i^a = ^i ■ x{k)i^a~i + x{k)i^a (6.28) 

for all values of a. 

6.3.2 The 2-category EqFlag^ 

The 2-category EqFlag^ is the equivariant analog of Flag^. 

Definition 6.10. The additive k-linear 2-category EqFlag^ is the idempotent completion 
inside of Bim of the 2-category consisting of 

• objects: the graded rings H^ for all k = (/cq, ki, /c2, . . . , kn) with < ki < k2 ■ ■ ■ < 

kn = N. 

• morphisms: generated by the graded (if^,iJ^)-bimodule H^, the graded {H'^,i^,H^)- 
bimodules H^+i and the graded (if^,if'^^^)-bimodule H^+, for all i E I, together with 
their shifts Hg{t}, H^+^{t}, and H^+,{t} for t e Z. The bimodules H^ = H^{0} are 
the identity 1-morphisms. Thus, a morphism from H^ to if '^ is a finite direct sum 
of graded {H'^g,, iJ^)-bimodules of the form 

for signed sequences j = SiS2 ■ ■ ■ Sm with ix = j x ^ ^■ 

• 2-morphisms: degree-preserving bimodule maps. 

There is a graded additive subcategory EqFlag^ of Bim* with the same objects and 
1-morphisms as EqFlag^, and 

EqFlag;,(Mi,M2) := EqFlag^(Mi{t}, M2). (6.29) 



Equivariant representation 81 

6.3.3 Equivariant representation F^ 

A 2-representation of the 2-category U*{sl2) = l^Zi^h) was constructed in [22] using equiv- 
ariant coliomology of partial flag varieties. Here we extend that construction to the 5[„-case 
and define a 2-representation F^: U^ -^ EqFlag^. 

The verification that Fjv is 2-representation with the assignments given in Section 16.11 
used only the relations fl5.2ip and fl5.25p - fl5.33p . together with Lemma [5^ and Corollary l5.5l 
which both follow from these relations. 

To define the equivariant 2-representation F^ ordinary cohomology rings are replaced 
by equivariant cohomology rings and bimodule maps associated to 2-morphisms are defined 
the same way, except that the dual generators x{k)j^a niust be redefined. Set x{k)j^o = 1 
and inductively define 

x{K)j,a = - XI x{k)jjx{K)j,a-f- (6.30) 

For example, 

x{,^j,i = -xik)j^i 

x{k)j^2 = -x{k)j^ix{k)j^i - x{k)j^2 = ^ikYjA - x{k)j,2 

xik)j,3 = -a;(^)j,i + 2 x{k)j^i x{k)j^2 - x{k)j,-i 

In the nonequivariant cohomology rings the two definitions (15.51) and fl6.30p of dual gener- 
ators agree, but not in the equivariant cohomology ring. 

Remark 6.11. We could have started with the definition oix{k)j^a given in f l6.30p and used 
this definition to construct the non-equivariant representation in the previous section. After 
all, the two definitions (15. 5p and fl6.30p are equivalent in the non-equivariant cohomology 
ring. However, the recursive definition f l6.30p makes it more cumbersome to calculate with 
in practice and that is why we used (15. 5p to construct the non-equivariant representation. 

Lemma 6.12. With ^(fc)^-^ redefined as in flOOD the relations flOT]) and fl5:25|) - fl5:33|l 


Equivariant representation 82 

hold in the equivariant cohomology rings: 

^X{k)jjx{k)j^a-f = Safi (6.31) 


x{k)j^a = x{+ik)j^a ioi J ^i,i + I (6.32) 

x{k)i+i^a = 6 ■ x{+ik)i+i^a-l + x{+ik)i+i^a (6.33) 

X[+ik)i^a = <,i ■ X[k_)i^a,-1 + ^^(kJj^q (6.34) 


x(+,fc)i+i,„ = 5]](-l)-^C^x(^),+i,„_/ (6.35) 


a;(^)i,a = ^x{+ik)i^a-f^! (6.36) 


-1)"C = ^x{+ik)i+i^a-f X{k)i+ij = ^x{+ik)ia-gX{k)i^g 

/=o s=o 


a^(^)j,a = x{+ik)j^a iij^i,i + l (6.38) 

3^(+j^)j,a-i M + 3;(_(_jA;)j_Q, II J — z 

2;(fc),-,a = { a (6.39) 

E{-lrx{+,k)^+l,a-f^I ifj=2 + l 


E (-l)°x(A;),„_/ e/ if J 



a;(+.^),,a = <( ' (6.40) 

a;(^)j+i,a-i ^i + x{k)i+i^a a j = i + l 

Proof. Equation (16.311) follows from the definition fl6.30p of the dual elements x{k)j^a- 
Equations (I6.32l) -( l6.34l) equate two images of noncanonical generators under the inclusions 
( 106|) and (lOTI) . Equations (lOSj) and (Km follow from (I03D and (lOil) . The first 
equality in (16.371) is proven as follows: 

a a 01— f 

(-i)"5^x(+,fc),+i,«_;^(i)^ ^ $^E(-i)°^'^'^(^)^+i.-/-3^(KZ7 

/=0 /=0 3=0 

a OL—g 

5=0 /=0 

i(OT1 i 

E(-l)"^'C''^a-.,o = it- (6.41) 



Equivariant representation 83 

The second equality above is just a re- indexing of the summation. The second equation in 
fl6.37p is proven similarly. 

Equation ( I6.38P follows from (16.32^ and the definition of x{k)j^a- Equations (16.391) and 
fjODjl follow from flg:^ and flOIj) and the definition of x{^)~. D 

Theorem 6.13. F^: U^ -^ EqFlag^ is a 2-representation. 

Proof. It is clear that F^ preserves the degree associated to 2-morphisms since the 2- 
representation Fat does. The proof that F^ preserves the relations in U^ can be copied 
line by line from the proof the Fat preserves the relations. By Lemma [6. 121 all identities used 
in the proof of Theorem 16.91 hold in EqFlag^ with the dual elements redefined according 
to (I63U|) . D 

Theorem 6.14. The 2-representations Fjv and F^ categorify the irreducible representation 

Vn of U(5[„) with highest weight (A^, 0, . . . , 0). 

Proof. Idempotent bimodule maps split in Bim, so by the universal property of the 
Karoubi envelope the additive 2-functors Fat: U — > Flag^ and F^: U -^ EqFlag^, 
obtained using the isomorphism Tj: U* ^ U^ and restricting to degree preserving 2- 
morphisms, extend to 2-representations of U. 

The rings Hk and H^ are graded local rings so that every projective module is free, and 
they both have (up to isomorphism and grading shift) a unique graded indecomposable 
projective module. The Grothendieck group of the category 0^ Hk—gmod (respectively 
0^, if^— gmod) is a free Z[g, g~^]-module with basis elements [Hk\ (respectively [H^]) over 
all A;, where g* acts by shifting the grading degree by i. Thus, we have 

^0 ( ^f -gmod) = i^o ( i^i-gmod) ^ ^V^ (6.42) 

k k 

as Z[g, g~^]-modules, where the sums are over all sequences < fci < ■ ■ ■ < A;„ = A^, and 
jXn is a representation of yiU(s[n), an integral form of the representation Vn of U(5l„). 

The bimodules F(1a), T{£j^iX\) and F(£^_j1a) (or equivalently F*^(1a), V^i^E+ilx) and 
T'^[8-il\)) induce functors on the graded module categories given by tensoring with these 
bimodules. The functors 



Hk ®Hk_ - '■ Hk-gmod -^ Hk-gmod (6.43) 

Hk+^ ^Hk - {1 + ki-i - ki+i}: iJfc-gmod -> iJ_^,fc-gmod (6.44) 
Hf.+^ Or ,^ - {-ki} : H^^k-gmod -^ Hk-gmod (6.45) 


have both left and right adjoints and commute with the shift functor, so they induce 
1j[q, g~^]-module maps on Grothendieck groups. Furthermore, the 2-functor Tjy respects 
the relations of W, so by Propositions 13 . 241 - 13 . 261 these functors satisfy relations lifting those 
of U. D 


Nondegeneracy ofU^{sln 


6.4 Nondegeneracy of W^(s[n) 

Lemma 6.15. The surjective graded k-algebra homomorphism 


of Proposition 13.61 is an isomorphism. 


Proof. Injectivity is estabhshed by showing that for each M e N there exists some large 
N such that the images of bubble monomials in W^(1a, 1a) in variables of degree less than 
M act by linearly independent operators on H^ for A = \{k), under the 2-functor T%. By 
a direct calculation 





\ Ai-l + Q / 

(-1)" Yl X{k)i,a-f X{k),+ij, 









as bimodule endomorphisms of H^. After expanding the x{k)j^a using (16.301) we have 


\ Xi-l+a / 
( ^ \ 


\ -A,-l+/3 / 

(-l)°(x(A;)i+i,, - x{k),,a) + (products of lower ] 

V order terms / 

-^Tixik^f, - x(A;),+i,^) + ("products of lower"] . 

V order terms / 

Only one orientation for dotted bubbles labelled by vertex i is allowed in a bubble monomial 
in IIa, see (I3.24p . The image under F*^ of a bubble monomial is composed of products of 
bimodule maps of the above form where, in the equivariant cohomology ring, sums of 
products of element of the form {x{k)i+i^a — x{k)i^a), respectively {x{k)i^i3 — x(A;)j+i,/3) 
are independent provided a,P < hi — ki^i and a,/5 < fc^+i — hi so that (a;(fc)i+i,a — 
3^(^)i+i,o)) respectively (a;(fc)j^/3 — x{k)i^i^p), is nonzero. By taking A^ large, we can ensure 
this condition is satisfied for any fixed M . Hence, images of bubble monomials in W^(1a, 1a) 
are independent. D 

Lemma 6.16. There is an isomorphism of graded k-algebras 

l'^:R{v)®TIx -^ Uti£ulx,£ulx) 
D®D^ ^ LxiD).D^, 

with Lx given by (I4.18P and D^^ a bubble monomial in IIa. 



Nondegeneracy ofU^{sln) 85 

Proof. Surjectivity of l\ follows from Lemma [3l9] and the isomorphism E from Section l4.2.1[ 
Injectivity of l\ is established by showing that for each M G N there exists some large 
N such that degree M elements lx{D) . D^,, as D and D^ run over a basis of -R(i^), 
respectively IIa, act by linear independent operators under the 2- representation F^. The 
2-functor F*^ must map horizontal composites to horizontal composites (given by tensor 
products) in EqFlag^, so that 

F^(.;,(D).D^) = F^(.,(D))®^gF^(DO := fD(»HogD^ (6.49) 

for bimodule maps go^ '■ H^ — ^ H'j^ and fo '■ H9i — > H?j for some i, j G Seq(i/). 




For large enough A^, the bimodule T'^^Si.lx) contains VoiuiO ^s a subspace. From the 
definition of lx (see (14.181) ). together with the definitions of the bimodule maps associated 
to R{i^) generators (see Section [B.1.2p . it is clear that the action on Vo£^{^) given by the 2- 
functor F"^ coincides with the action oi R{i>) on Voiu defined in ^E\. In particular, bimodule 
maps fo corresponding to basis elements of -R(z^) must act by linear independent operators 
(see [m proof of Theorem 2.5]) on the subspace Voi-u{^) of F*^(£^i^1a). Furthermore, from 
the definitions of bimodule maps associated to R{v) generators it is also clear that f^ fixes 
all other generators of V^i^Eylx). 

By Lemma 16. 15^ for large N the bimodule maps g^^ act by linearly independent oper- 

ators on H^ . Write 



(j,A)-l + Q 



for (i. A) > 0, 
for (i. A) < 0. 


and let D^, = iii^^aiT^ 1,2,02 ' ' ' ^£r,ar- Then D^, acts on i/^ = H'^i® ^a H^ via F'^(Idi^ . Dt,) 
1 ®fjG gr,^. From (16.471) we have 

1®H^^D,: 101 ^ ^x{k)i^+ij^x{k)i,^^^^f^ J ■■■ ^x{k)i^+ij^x{k)i^^^^_f^ j (6.51) 
\/i=0 / \/,=o 

in H'fi. After expanding the x{k)i,a using their definition (I6.30p . we have an expression for 
the action of g^,^ on H'fi strictly in terms of variables x{k)i,a- 

The bimodule maps fo ®h'^ go^ are linearly independent operators since the bimod- 
ule maps /d and gn^ are separately independent and act on algebraically independent 




generators of H^i : 





l/3i 1/32 1/3 

■~ f •■■ t 

il J2 jm 



)^^'", and the 

where the sum over the /3a 's is determined by the action of fo on ^° 
sum over the x{k)ej^^^^, represented by labelled boxes on the far right, is determined from 
f l6.5ip by expanding the x{k)-. In particular, f^ fixes variables x{k)e^a represented by 
labelled boxes on the far right and g^ acts only on such boxes. D 

This concludes the proof of Theorem 11.31 stated in the introduction, so that 



is an isomorphism. Proposition 13.281 allows us to view 2-functors ip, uj, a, and r as categori- 
fications of symmetries ip, ^, o", and r of _4U(s[„). The graded homs between 1-morphisms 
in U{sln) categorify the semilinear form (, )^ given by fl3.11Up . 

Simple nondegeneracy observations. We can assume that k is only a commutative 
ring, work over this ring from the start, and say that the calculus is nondegenerate over k 
if U*{Silx,Sjlx) is a free k-module with a basis Bijx for all j, i and A. We don't know 
any examples of a root datum and ring k when the calculus is degenerate. Over a field 
the nondegeneracy of the calculus depends only on the root datum and the characteristic 
of k. If the calculus is nondegenerate over (Q, it is nondegenerate over Z and over any 
commutative ring k. 

Multigrading. The multigrading introduced at the end of [IT] on rings -R(i^) extends 
to a multigrading on each hom space U*{Silx,£jlx). The Karoubian envelope of the 
corresponding multigraded 2-category should categorify the multi-parameter deformation 


[1] S. Ariki. On the decomposition numbers of the Hecke algebra of G{m, 1, n). J. Math. 
Kyoto Umv., 36(4):789-808, 1996. 

[2] A. Beilinson, G. Lusztig, and R. MacPherson. A geometric setting for the quantum 
deformation of GL„. Duke Math. J., 61(2):655-677, 1990. 

[3] J. Bernstein, I. B. Frenkel, and M. Khovanov. A categorification of the Temperley- 
Lieb algebra and Schur quotients of U(sl(2)) via projective and Zuckerman functors. 
Selecta Math. (N.S.), 5(2):199-241, 1999, |math.Q A/0002087[ 



[4] F. Borceux. Handbook of categorical algebra. 1, volume 50 of Encyclopedia of Mathe- 
matics and its Applications. Cambridge University Press, Cambridge, 1994. 

[5] S. Cautis and J. Kamnitzer. Knot homology via derived categories of coherent sheaves. 
II, sl(m) case, 2007, math.AG/0710.3216. 

[6] S. Cautis and J. Kamnitzer. Knot homology via derived categories of coherent sheaves. 

I, sl(2)-case. Duke Math. J., 142(3):511-588, 2008, math. AG/0701194 

[7] J. Chuang and R. Rouquier. Derived equivalences for symmetric groups and sL2- 
categorification. Ann. of Math., 167:245-298, 2008, math.RT/0407205 . 

L. Crane and I. B. Frenkel. Four-dimensional topological quantum field theory, 
Hopf categories, and the canonical bases. J. Math. Phys., 35(10):5136-5154, 1994, 
|hep -th/9405183 ] 

[9] V. Drinfeld. Quantum groups. In Proceedings of the International Congress of Math- 
ematicians, Vol. 1, 2 (Berkeley, Calif., 1986), pages 798-820, Providence, RI, 1987. 
Amer. Math. Soc. 

[10] I. B. Frenkel, M. Khovanov, and C. Stroppel. A categorification of finite-dimensional 
irreducible representations of quantum sl(2) and their tensor products. Selecta Math. 
(N.S.), 12(3-4) :379-431, 2006,|math.QA/0511467[ 

[11] W. Fulton. Equivariant cohomology in algebraic geometry. Eilenberg 

lectures, Columbia University. Notes by Dave Anderson, available at 
http : //www . math . Isa . umich . edu/~dandersn/eilenberg/, 2007. 

[12] N. Canter and M. Kapranov. Representation and character theory in 2-categories. 

Adv. Math., 217(5):2268-2300, 2008, |math.KT/0602510 

[13] M. Jimbo. A g-difference analogue of U{g) and the Yang-Baxter equation. Lett. Math. 
Phys., 10(l):63-69, 1985. 

[14] M. Kashiwara. Global crystal bases of quantum groups. Duke Math. J., 69(2):455-485, 

[15] M. Khovanov. A functor- valued invariant of tangles. Algebr. Ceom. Toj^o/., 2:665-741 
(electronic), 2002, ,math.QA/0103190 . 

[16] M. Khovanov and A. Lauda. A diagrammatic approach to categorification of quantum 
groups I, 2008, math.QA/0803.4121. 

[17] M. Khovanov and A. Lauda. A diagrammatic approach to categorification of quantum 
groups II, 2008, math.QA/0804.2080. 

[18] M. Khovanov and L. Rozansky. Matrix factorizations and link homology. Fund. Math., 

199(1):1-91, 2008, math.QA/0401268 














A. Kleshchev. Linear and projective representations of symmetric groups, volume 163 
of Cambridge Tracts in Mathematics. Cambridge U. Press, 2005. 

P. Kronheimer and H. Nakajima. Yang-Mills instantons on ALE gravitational instan- 
tons. Math. Ann., 288(2):263-307, 1990. 

A. D. Lauda. A categorification of quantum sl(2), 2008, math. Q A/0803. 3652. 

A. D. Lauda. Categorified quantum sl(2) and equivariant coliomology of iterated flag 
varieties, 2008, math.QA/0803.3848. 

G. Lusztig. Canonical bases in tensor products. Proc. Nat. Acad. Sci. U.S.A., 
89(17):8177-8179, 1992. 

G. Lusztig. Introduction to quantum groups, volume 110 of Progress in Mathematics. 
Birkhauser Boston Inc., Boston, MA, 1993. 

G. Lusztig. Canonical bases and Hall algebras. In Representation theories and algebraic 
geometry (Montreal, PQ, 1997), volume 514 of NATO Adv. Sci. Inst. Ser. C Math. 
Phys. Sci., pages 365-399. Kluwer Acad. Publ., Dordrecht, 1998. 

H. Nakajima. Instantons on ALE spaces, quiver varieties, and Kac-Moody algebras. 
Duke Math. J., 76(2):365-416, 1994. 

M. Reineke. Monomials in canonical bases of quantum groups and quadratic forms. 
J. Pure Appl. Algebra, 157(2-3):301-309, 2001. 

J. Rosenberg. Algebraic K-theory and its applications, volume 147 of Graduate Texts 
in Mathematics. Springer- Verlag, New York, 1994. 

R. Rouquier. Higher representation theory. Talk at IAS, March 2008. 

R. Rouquier. Higher representations of Kac-Moody algebras. Work in progress. 

J. Sussan. Category O and sl(k) link invariants, 2007, math.QA/0701045 

C. Weibel. The K-book: An introduction to algebraic K-theory. available online, 
http : //www . math . rutgers . edu/~weibel/Kbook . html. 

[33] H. Zheng. Categorification of integrable representations of quantum groups, 2008, 

M.K.: Department oi Mathematics, Columbia University, New York, NY 10027 
email : khovanovOmath . Columbia . edu 

A.L: Department of Mathematics, Columbia University, New York, NY 10027